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 +) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
16 additional bytes when new memory had to be allocated, not just
17 when reusing an existing buffer.
20 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
21 instead. BIO_gethostbyname() does not know what timeouts are
22 appropriate, so entries would stay in cache even when they have
24 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
26 +) New command line and configuration option 'utf8' for the req command.
27 This allows field values to be specified as UTF8 strings.
30 +) Add -multi and -mr options to "openssl speed" - giving multiple parallel
31 runs for the former and machine-readable output for the latter.
34 +) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
35 of the e-mail address in the DN (i.e., it will go into a certificate
36 extension only). The new configuration file option 'email_in_dn = no'
38 [Massimiliano Pala madwolf@openca.org]
40 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
41 faced with a pathologically small ClientHello fragment that does
42 not contain client_version: Instead of aborting with an error,
43 simply choose the highest available protocol version (i.e.,
44 TLS 1.0 unless it is disabled). In practice, ClientHello
45 messages are never sent like this, but this change gives us
46 strictly correct behaviour at least for TLS.
49 +) Change all functions with names starting with des_ to be starting
50 with DES_ instead. This because there are increasing clashes with
51 libdes and other des libraries that are currently used by other
52 projects. The old libdes interface is provided, as well as crypt(),
53 if openssl/des_old.h is included. Note that crypt() is no longer
54 declared in openssl/des.h.
56 NOTE: This is a major break of an old API into a new one. Software
57 authors are encouraged to switch to the DES_ style functions. Some
58 time in the future, des_old.h and the libdes compatibility functions
59 will be completely removed.
62 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
63 never resets s->method to s->ctx->method when called from within
64 one of the SSL handshake functions.
65 [Bodo Moeller; problem pointed out by Niko Baric]
67 +) Test for certificates which contain unsupported critical extensions.
68 If such a certificate is found during a verify operation it is
69 rejected by default: this behaviour can be overridden by either
70 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
71 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
72 X509_supported_extension() has also been added which returns 1 if a
73 particular extension is supported.
76 +) New functions/macros
78 SSL_CTX_set_msg_callback(ctx, cb)
79 SSL_CTX_set_msg_callback_arg(ctx, arg)
80 SSL_set_msg_callback(ssl, cb)
81 SSL_set_msg_callback_arg(ssl, arg)
83 to request calling a callback function
85 void cb(int write_p, int version, int content_type,
86 const void *buf, size_t len, SSL *ssl, void *arg)
88 whenever a protocol message has been completely received
89 (write_p == 0) or sent (write_p == 1). Here 'version' is the
90 protocol version according to which the SSL library interprets
91 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
92 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
93 the content type as defined in the SSL 3.0/TLS 1.0 protocol
94 specification (change_cipher_spec(20), alert(21), handshake(22)).
95 'buf' and 'len' point to the actual message, 'ssl' to the
96 SSL object, and 'arg' is the application-defined value set by
97 SSL[_CTX]_set_msg_callback_arg().
99 'openssl s_client' and 'openssl s_server' have new '-msg' options
100 to enable a callback that displays all protocol messages.
102 TODO: SSL 2.0, doc/ssl/, doc/apps/
105 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
106 (sent using the client's version number) if client_version is
107 smaller than the protocol version in use. Also change
108 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
109 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
110 the client will at least see that alert.
113 +) Modify the behaviour of EVP cipher functions in similar way to digests
114 to retain compatibility with existing code.
117 +) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
118 compatibility with existing code. In particular the 'ctx' parameter does
119 not have to be to be initialized before the call to EVP_DigestInit() and
120 it is tidied up after a call to EVP_DigestFinal(). New function
121 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
122 EVP_MD_CTX_copy() changed to not require the destination to be
123 initialized valid and new function EVP_MD_CTX_copy_ex() added which
124 requires the destination to be valid.
126 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
127 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
130 +) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
131 so that complete 'Handshake' protocol structures are kept in memory
132 instead of overwriting 'msg_type' and 'length' with 'body' data.
135 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
139 +) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
140 [Massimo Santin via Richard Levitte]
142 +) Major restructuring to the underlying ENGINE code. This includes
143 reduction of linker bloat, separation of pure "ENGINE" manipulation
144 (initialisation, etc) from functionality dealing with implementations
145 of specific crypto iterfaces. This change also introduces integrated
146 support for symmetric ciphers and digest implementations - so ENGINEs
147 can now accelerate these by providing EVP_CIPHER and EVP_MD
148 implementations of their own. This is detailed in crypto/engine/README
149 as it couldn't be adequately described here. However, there are a few
150 API changes worth noting - some RSA, DSA, DH, and RAND functions that
151 were changed in the original introduction of ENGINE code have now
152 reverted back - the hooking from this code to ENGINE is now a good
153 deal more passive and at run-time, operations deal directly with
154 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
155 dereferencing through an ENGINE pointer any more. Also, the ENGINE
156 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
157 they were not being used by the framework as there is no concept of a
158 BIGNUM_METHOD and they could not be generalised to the new
159 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
160 ENGINE_cpy() has been removed as it cannot be consistently defined in
164 +) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
167 +) Change mkdef.pl to sort symbols that get the same entry number,
168 and make sure the automatically generated functions ERR_load_*
169 become part of libeay.num as well.
172 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
173 client receives HelloRequest while in a handshake.
174 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
176 +) New function SSL_renegotiate_pending(). This returns true once
177 renegotiation has been requested (either SSL_renegotiate() call
178 or HelloRequest/ClientHello receveived from the peer) and becomes
179 false once a handshake has been completed.
180 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
181 sends a HelloRequest, but does not ensure that a handshake takes
182 place. SSL_renegotiate_pending() is useful for checking if the
183 client has followed the request.)
186 +) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
187 By default, clients may request session resumption even during
188 renegotiation (if session ID contexts permit); with this option,
189 session resumption is possible only in the first handshake.
192 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
193 should end in 'break', not 'goto end' which circuments various
194 cleanups done in state SSL_ST_OK. But session related stuff
195 must be disabled for SSL_ST_OK in the case that we just sent a
198 Also avoid some overhead by not calling ssl_init_wbio_buffer()
199 before just sending a HelloRequest.
200 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
202 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
203 reveal whether illegal block cipher padding was found or a MAC
204 verification error occured. (Neither SSLerr() codes nor alerts
205 are directly visible to potential attackers, but the information
206 may leak via logfiles.)
208 Similar changes are not required for the SSL 2.0 implementation
209 because the number of padding bytes is sent in clear for SSL 2.0,
210 and the extra bytes are just ignored. However ssl/s2_pkt.c
211 failed to verify that the purported number of padding bytes is in
215 +) Add some demos for certificate and certificate request creation.
218 +) Make maximum certificate chain size accepted from the peer application
219 settable (SSL*_get/set_max_cert_list()), as proposed by
220 "Douglas E. Engert" <deengert@anl.gov>.
223 +) Add support for shared libraries for Unixware-7 and support including
224 shared libraries for OpenUNIX-8 (Boyd Lynn Gerber <gerberb@zenez.com>).
227 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
228 'wristwatch attack' using huge encoding parameters (cf.
229 James H. Manger's CRYPTO 2001 paper). Note that the
230 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
231 encoding parameters and hence was not vulnerable.
234 +) Add a "destroy" handler to ENGINEs that allows structural cleanup to
235 be done prior to destruction. Use this to unload error strings from
236 ENGINEs that load their own error strings. NB: This adds two new API
237 functions to "get" and "set" this destroy handler in an ENGINE.
240 +) Alter all existing ENGINE implementations (except "openssl" and
241 "openbsd") to dynamically instantiate their own error strings. This
242 makes them more flexible to be built both as statically-linked ENGINEs
243 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
244 Also, add stub code to each that makes building them as self-contained
245 shared-libraries easier (see README.ENGINE).
248 +) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
249 implementations into applications that are completely implemented in
250 self-contained shared-libraries. The "dynamic" ENGINE exposes control
251 commands that can be used to configure what shared-library to load and
252 to control aspects of the way it is handled. Also, made an update to
253 the README.ENGINE file that brings its information up-to-date and
254 provides some information and instructions on the "dynamic" ENGINE
255 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
259 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
261 *) Make it possible to unload ranges of ERR strings with a new
262 "ERR_unload_strings" function.
265 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
266 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
267 followed by modular reduction.
268 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
270 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
271 equivalent based on BN_pseudo_rand() instead of BN_rand().
274 +) Add a copy() function to EVP_MD.
277 +) Make EVP_MD routines take a context pointer instead of just the
278 md_data void pointer.
281 +) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
282 that the digest can only process a single chunk of data
283 (typically because it is provided by a piece of
284 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
285 is only going to provide a single chunk of data, and hence the
286 framework needn't accumulate the data for oneshot drivers.
289 +) As with "ERR", make it possible to replace the underlying "ex_data"
290 functions. This change also alters the storage and management of global
291 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
292 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
293 index counters. The API functions that use this state have been changed
294 to take a "class_index" rather than pointers to the class's local STACK
295 and counter, and there is now an API function to dynamically create new
296 classes. This centralisation allows us to (a) plug a lot of the
297 thread-safety problems that existed, and (b) makes it possible to clean
298 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
299 such data would previously have always leaked in application code and
300 workarounds were in place to make the memory debugging turn a blind eye
301 to it. Application code that doesn't use this new function will still
302 leak as before, but their memory debugging output will announce it now
303 rather than letting it slide.
305 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
306 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
307 has a return value to indicate success or failure.
310 +) Make it possible to replace the underlying "ERR" functions such that the
311 global state (2 LHASH tables and 2 locks) is only used by the "default"
312 implementation. This change also adds two functions to "get" and "set"
313 the implementation prior to it being automatically set the first time
314 any other ERR function takes place. Ie. an application can call "get",
315 pass the return value to a module it has just loaded, and that module
316 can call its own "set" function using that value. This means the
317 module's "ERR" operations will use (and modify) the error state in the
318 application and not in its own statically linked copy of OpenSSL code.
321 +) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
322 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
323 the operation, and provides a more encapsulated way for external code
324 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
325 to use these functions rather than manually incrementing the counts.
327 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
330 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
331 This function was broken, as the check for a new client hello message
332 to handle SGC did not allow these large messages.
333 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
336 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
339 +) Add EVP test program.
342 +) Add symmetric cipher support to ENGINE. Expect the API to change!
345 +) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
346 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
347 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
348 These allow a CRL to be built without having to access X509_CRL fields
349 directly. Modify 'ca' application to use new functions.
352 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
353 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
356 *) Rework the configuration and shared library support for Tru64 Unix.
357 The configuration part makes use of modern compiler features and
358 still retains old compiler behavior for those that run older versions
359 of the OS. The shared library support part includes a variant that
360 uses the RPATH feature, and is available through the special
361 configuration target "alpha-cc-rpath", which will never be selected
363 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
365 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
366 with the same message size as in ssl3_get_certificate_request().
367 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
368 messages might inadvertently be reject as too long.
369 [Petr Lampa <lampa@fee.vutbr.cz>]
371 +) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
372 bug workarounds. Rollback attack detection is a security feature.
373 The problem will only arise on OpenSSL servers when TLSv1 is not
374 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
375 Software authors not wanting to support TLSv1 will have special reasons
376 for their choice and can explicitly enable this option.
377 [Bodo Moeller, Lutz Jaenicke]
379 +) Rationalise EVP so it can be extended: don't include a union of
380 cipher/digest structures, add init/cleanup functions. This also reduces
381 the number of header dependencies.
384 +) Make DES key schedule conform to the usual scheme, as well as
385 correcting its structure. This means that calls to DES functions
386 now have to pass a pointer to a des_key_schedule instead of a
387 plain des_key_schedule (which was actually always a pointer
391 +) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
394 *) Modified SSL library such that the verify_callback that has been set
395 specificly for an SSL object with SSL_set_verify() is actually being
396 used. Before the change, a verify_callback set with this function was
397 ignored and the verify_callback() set in the SSL_CTX at the time of
398 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
399 to allow the necessary settings.
402 +) Initial reduction of linker bloat: the use of some functions, such as
403 PEM causes large amounts of unused functions to be linked in due to
404 poor organisation. For example pem_all.c contains every PEM function
405 which has a knock on effect of linking in large amounts of (unused)
406 ASN1 code. Grouping together similar functions and splitting unrelated
407 functions prevents this.
410 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
411 explicitly to NULL, as at least on Solaris 8 this seems not always to be
412 done automatically (in contradiction to the requirements of the C
413 standard). This made problems when used from OpenSSH.
416 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
417 dh->length and always used
419 BN_rand_range(priv_key, dh->p).
421 BN_rand_range() is not necessary for Diffie-Hellman, and this
422 specific range makes Diffie-Hellman unnecessarily inefficient if
423 dh->length (recommended exponent length) is much smaller than the
424 length of dh->p. We could use BN_rand_range() if the order of
425 the subgroup was stored in the DH structure, but we only have
430 BN_rand(priv_key, l, ...)
432 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
438 RSA_eay_public_encrypt
439 RSA_eay_private_decrypt
440 RSA_eay_private_encrypt (signing)
441 RSA_eay_public_decrypt (signature verification)
443 (default implementations for RSA_public_encrypt,
444 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
445 always reject numbers >= n.
448 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
449 to synchronize access to 'locking_thread'. This is necessary on
450 systems where access to 'locking_thread' (an 'unsigned long'
451 variable) is not atomic.
454 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
455 *before* setting the 'crypto_lock_rand' flag. The previous code had
456 a race condition if 0 is a valid thread ID.
457 [Travis Vitek <vitek@roguewave.com>]
459 +) Cleanup of EVP macros.
462 +) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
466 +) Add initial OCSP responder support to ocsp application. The
467 revocation information is handled using the text based index
468 use by the ca application. The responder can either handle
469 requests generated internally, supplied in files (for example
470 via a CGI script) or using an internal minimal server.
473 +) Add configuration choices to get zlib compression for TLS.
476 +) Changes to Kerberos SSL for RFC 2712 compliance:
477 1. Implemented real KerberosWrapper, instead of just using
478 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
479 2. Implemented optional authenticator field of KerberosWrapper.
481 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
482 and authenticator structs; see crypto/krb5/.
484 Generalized Kerberos calls to support multiple Kerberos libraries.
485 [Vern Staats <staatsvr@asc.hpc.mil>,
486 Jeffrey Altman <jaltman@columbia.edu>
489 +) Cause 'openssl speed' to use fully hard-coded DSA keys as it
490 already does with RSA. testdsa.h now has 'priv_key/pub_key'
491 values for each of the key sizes rather than having just
492 parameters (and 'speed' generating keys each time).
495 -) OpenSSL 0.9.6b released [9 July 2001]
497 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
498 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
499 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
500 PRNG state recovery was possible based on the output of
501 one PRNG request appropriately sized to gain knowledge on
502 'md' followed by enough consecutive 1-byte PRNG requests
503 to traverse all of 'state'.
505 1. When updating 'md_local' (the current thread's copy of 'md')
506 during PRNG output generation, hash all of the previous
507 'md_local' value, not just the half used for PRNG output.
509 2. Make the number of bytes from 'state' included into the hash
510 independent from the number of PRNG bytes requested.
512 The first measure alone would be sufficient to avoid
513 Markku-Juhani's attack. (Actually it had never occurred
514 to me that the half of 'md_local' used for chaining was the
515 half from which PRNG output bytes were taken -- I had always
516 assumed that the secret half would be used.) The second
517 measure makes sure that additional data from 'state' is never
518 mixed into 'md_local' in small portions; this heuristically
519 further strengthens the PRNG.
522 +) Speed up EVP routines.
525 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
526 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
527 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
528 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
530 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
531 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
532 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
535 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
537 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
540 *) Fix crypto/bn/asm/mips3.s.
543 *) When only the key is given to "enc", the IV is undefined. Print out
544 an error message in this case.
547 +) Added the OS2-EMX target.
548 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
550 +) Rewrite apps to use NCONF routines instead of the old CONF. New functions
551 to support NCONF routines in extension code. New function CONF_set_nconf()
552 to allow functions which take an NCONF to also handle the old LHASH
553 structure: this means that the old CONF compatible routines can be
554 retained (in particular wrt extensions) without having to duplicate the
555 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
558 *) Handle special case when X509_NAME is empty in X509 printing routines.
561 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
562 positive and less than q.
565 +) Enhance the general user interface with mechanisms for inner control
566 and with possibilities to have yes/no kind of prompts.
569 +) Change all calls to low level digest routines in the library and
570 applications to use EVP. Add missing calls to HMAC_cleanup() and
571 don't assume HMAC_CTX can be copied using memcpy().
572 [Verdon Walker <VWalker@novell.com>, Steve Henson]
574 +) Add the possibility to control engines through control names but with
575 arbitrary arguments instead of just a string.
576 Change the key loaders to take a UI_METHOD instead of a callback
577 function pointer. NOTE: this breaks binary compatibility with earlier
578 versions of OpenSSL [engine].
579 Adapt the nCipher code for these new conditions and add a card insertion
583 +) Enhance the general user interface with mechanisms to better support
584 dialog box interfaces, application-defined prompts, the possibility
585 to use defaults (for example default passwords from somewhere else)
586 and interrupts/cancellations.
589 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
590 used: it isn't thread safe and the add_lock_callback should handle
592 [Paul Rose <Paul.Rose@bridge.com>]
594 *) Verify that incoming data obeys the block size in
595 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
598 +) Tidy up PKCS#12 attribute handling. Add support for the CSP name
599 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
603 [Ulf Möller, Bodo Möller]
605 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
606 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
607 when fixing the server behaviour for backwards-compatible 'client
608 hello' messages. (Note that the attack is impractical against
609 SSL 3.0 and TLS 1.0 anyway because length and version checking
610 means that the probability of guessing a valid ciphertext is
611 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
614 Before 0.9.5, the countermeasure (hide the error by generating a
615 random 'decryption result') did not work properly because
616 ERR_clear_error() was missing, meaning that SSL_get_error() would
617 detect the supposedly ignored error.
619 Both problems are now fixed.
622 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
623 (previously it was 1024).
626 +) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
627 tidy up some unnecessarily weird code in 'sk_new()').
628 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
630 +) Change the key loading routines for ENGINEs to use the same kind
631 callback (pem_password_cb) as all other routines that need this
635 *) Fix for compatibility mode trust settings: ignore trust settings
636 unless some valid trust or reject settings are present.
639 *) Fix for blowfish EVP: its a variable length cipher.
642 +) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
643 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
644 than this minimum value is recommended.
647 +) New random seeder for OpenVMS, using the system process statistics
648 that are easily reachable.
651 +) Windows apparently can't transparently handle global
652 variables defined in DLLs. Initialisations such as:
654 const ASN1_ITEM *it = &ASN1_INTEGER_it;
656 wont compile. This is used by the any applications that need to
657 declare their own ASN1 modules. This was fixed by adding the option
658 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
659 needed for static libraries under Win32.
662 +) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
663 setting of purpose and trust fields. New X509_STORE trust and
664 purpose functions and tidy up setting in other SSL functions.
667 +) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
668 structure. These are inherited by X509_STORE_CTX when it is
669 initialised. This allows various defaults to be set in the
670 X509_STORE structure (such as flags for CRL checking and custom
671 purpose or trust settings) for functions which only use X509_STORE_CTX
672 internally such as S/MIME.
674 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
675 trust settings if they are not set in X509_STORE. This allows X509_STORE
676 purposes and trust (in S/MIME for example) to override any set by default.
678 Add command line options for CRL checking to smime, s_client and s_server
682 +) Initial CRL based revocation checking. If the CRL checking flag(s)
683 are set then the CRL is looked up in the X509_STORE structure and
684 its validity and signature checked, then if the certificate is found
685 in the CRL the verify fails with a revoked error.
687 Various new CRL related callbacks added to X509_STORE_CTX structure.
689 Command line options added to 'verify' application to support this.
691 This needs some additional work, such as being able to handle multiple
692 CRLs with different times, extension based lookup (rather than just
693 by subject name) and ultimately more complete V2 CRL extension
697 +) Add a general user interface API (crypto/ui/). This is designed
698 to replace things like des_read_password and friends (backward
699 compatibility functions using this new API are provided).
700 The purpose is to remove prompting functions from the DES code
701 section as well as provide for prompting through dialog boxes in
702 a window system and the like.
705 *) Fix various bugs related to DSA S/MIME verification. Handle missing
706 parameters in DSA public key structures and return an error in the
707 DSA routines if parameters are absent.
710 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
711 in the current directory if neither $RANDFILE nor $HOME was set.
712 RAND_file_name() in 0.9.6a returned NULL in this case. This has
713 caused some confusion to Windows users who haven't defined $HOME.
714 Thus RAND_file_name() is changed again: e_os.h can define a
715 DEFAULT_HOME, which will be used if $HOME is not set.
716 For Windows, we use "C:"; on other platforms, we still require
717 environment variables.
719 +) Add "ex_data" support to ENGINE so implementations can add state at a
720 per-structure level rather than having to store it globally.
723 +) Make it possible for ENGINE structures to be copied when retrieved by
724 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
725 This causes the "original" ENGINE structure to act like a template,
726 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
727 operational state can be localised to each ENGINE structure, despite the
728 fact they all share the same "methods". New ENGINE structures returned in
729 this case have no functional references and the return value is the single
730 structural reference. This matches the single structural reference returned
731 by ENGINE_by_id() normally, when it is incremented on the pre-existing
735 +) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
736 needs to match any other type at all we need to manually clear the
740 +) Changes to the "openssl engine" utility to include;
741 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
742 about an ENGINE's available control commands.
743 - executing control commands from command line arguments using the
744 '-pre' and '-post' switches. '-post' is only used if '-t' is
745 specified and the ENGINE is successfully initialised. The syntax for
746 the individual commands are colon-separated, for example;
747 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
750 +) New dynamic control command support for ENGINEs. ENGINEs can now
751 declare their own commands (numbers), names (strings), descriptions,
752 and input types for run-time discovery by calling applications. A
753 subset of these commands are implicitly classed as "executable"
754 depending on their input type, and only these can be invoked through
755 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
756 can be based on user input, config files, etc). The distinction is
757 that "executable" commands cannot return anything other than a boolean
758 result and can only support numeric or string input, whereas some
759 discoverable commands may only be for direct use through
760 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
761 pointers, or other custom uses. The "executable" commands are to
762 support parameterisations of ENGINE behaviour that can be
763 unambiguously defined by ENGINEs and used consistently across any
764 OpenSSL-based application. Commands have been added to all the
765 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
766 control over shared-library paths without source code alterations.
769 +) Changed all ENGINE implementations to dynamically allocate their
770 ENGINEs rather than declaring them statically. Apart from this being
771 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
772 this also allows the implementations to compile without using the
773 internal engine_int.h header.
776 +) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
777 'const' value. Any code that should be able to modify a RAND_METHOD
778 should already have non-const pointers to it (ie. they should only
779 modify their own ones).
782 +) Made a variety of little tweaks to the ENGINE code.
783 - "atalla" and "ubsec" string definitions were moved from header files
784 to C code. "nuron" string definitions were placed in variables
785 rather than hard-coded - allowing parameterisation of these values
786 later on via ctrl() commands.
787 - Removed unused "#if 0"'d code.
788 - Fixed engine list iteration code so it uses ENGINE_free() to release
789 structural references.
790 - Constified the RAND_METHOD element of ENGINE structures.
791 - Constified various get/set functions as appropriate and added
792 missing functions (including a catch-all ENGINE_cpy that duplicates
793 all ENGINE values onto a new ENGINE except reference counts/state).
794 - Removed NULL parameter checks in get/set functions. Setting a method
795 or function to NULL is a way of cancelling out a previously set
796 value. Passing a NULL ENGINE parameter is just plain stupid anyway
797 and doesn't justify the extra error symbols and code.
798 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
799 flags from engine_int.h to engine.h.
800 - Changed prototypes for ENGINE handler functions (init(), finish(),
801 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
804 *) Move 'if (!initialized) RAND_poll()' into regions protected by
805 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
806 having multiple threads call RAND_poll() concurrently.
809 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
810 combination of a flag and a thread ID variable.
811 Otherwise while one thread is in ssleay_rand_bytes (which sets the
812 flag), *other* threads can enter ssleay_add_bytes without obeying
813 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
814 that they do not hold after the first thread unsets add_do_not_lock).
817 +) Implement binary inversion algorithm for BN_mod_inverse in addition
818 to the algorithm using long division. The binary algorithm can be
819 used only if the modulus is odd. On 32-bit systems, it is faster
820 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
821 roughly 5-15% for 256-bit moduli), so we use it only for moduli
822 up to 450 bits. In 64-bit environments, the binary algorithm
823 appears to be advantageous for much longer moduli; here we use it
824 for moduli up to 2048 bits.
827 *) Change bctest again: '-x' expressions are not available in all
831 -) OpenSSL 0.9.6a released [5 Apr 2001]
833 *) Fix a couple of memory leaks in PKCS7_dataDecode()
834 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
836 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
837 the default extension for executables, if any. Also, make the perl
838 scripts that use symlink() to test if it really exists and use "cp"
839 if it doesn't. All this made OpenSSL compilable and installable in
843 +) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
844 could not support the combine flag in choice fields.
847 -) Fix for asn1_GetSequence() for indefinite length constructed data.
848 If SEQUENCE is length is indefinite just set c->slen to the total
849 amount of data available.
850 [Steve Henson, reported by shige@FreeBSD.org]
851 [This change does not apply to 0.9.7.]
853 *) Change bctest to avoid here-documents inside command substitution
854 (workaround for FreeBSD /bin/sh bug).
855 For compatibility with Ultrix, avoid shell functions (introduced
856 in the bctest version that searches along $PATH).
859 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
860 with des_encrypt() defined on some operating systems, like Solaris
864 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
865 On the Importance of Eliminating Errors in Cryptographic
866 Computations, J. Cryptology 14 (2001) 2, 101-119,
867 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
870 *) MIPS assembler BIGNUM division bug fix.
873 *) Disabled incorrect Alpha assembler code.
876 -) Fix PKCS#7 decode routines so they correctly update the length
877 after reading an EOC for the EXPLICIT tag.
879 [This change does not apply to 0.9.7.]
881 *) Fix bug in PKCS#12 key generation routines. This was triggered
882 if a 3DES key was generated with a 0 initial byte. Include
883 PKCS12_BROKEN_KEYGEN compilation option to retain the old
884 (but broken) behaviour.
887 *) Enhance bctest to search for a working bc along $PATH and print
889 [Tim Rice <tim@multitalents.net> via Richard Levitte]
891 +) Add a 'copy_extensions' option to the 'ca' utility. This copies
892 extensions from a certificate request to the certificate.
895 +) Allow multiple 'certopt' and 'nameopt' options to be separated
896 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
897 file: this allows the display of the certificate about to be
898 signed to be customised, to allow certain fields to be included
899 or excluded and extension details. The old system didn't display
900 multicharacter strings properly, omitted fields not in the policy
901 and couldn't display additional details such as extensions.
904 *) Fix memory leaks in err.c: free err_data string if necessary;
905 don't write to the wrong index in ERR_set_error_data.
908 +) Function EC_POINTs_mul for simultaneous scalar multiplication
909 of an arbitrary number of elliptic curve points, optionally
910 including the generator defined for the EC_GROUP.
911 EC_POINT_mul is a simple wrapper function for the typical case
912 that the point list has just one item (besides the optional
916 +) First EC_METHODs for curves over GF(p):
918 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
919 operations and provides various method functions that can also
920 operate with faster implementations of modular arithmetic.
922 EC_GFp_mont_method() reuses most functions that are part of
923 EC_GFp_simple_method, but uses Montgomery arithmetic.
925 [Bodo Moeller; point addition and point doubling
926 implementation directly derived from source code provided by
927 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
929 +) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
932 Curves are EC_GROUP objects (with an optional group generator)
933 based on EC_METHODs that are built into the library.
935 Points are EC_POINT objects based on EC_GROUP objects.
937 Most of the framework would be able to handle curves over arbitrary
938 finite fields, but as there are no obvious types for fields other
939 than GF(p), some functions are limited to that for now.
942 +) Add the -HTTP option to s_server. It is similar to -WWW, but requires
943 that the file contains a complete HTTP response.
946 +) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
947 change the def and num file printf format specifier from "%-40sXXX"
948 to "%-39s XXX". The latter will always guarantee a space after the
949 field while the former will cause them to run together if the field
950 is 40 of more characters long.
953 +) Constify the cipher and digest 'method' functions and structures
954 and modify related functions to take constant EVP_MD and EVP_CIPHER
958 *) Implement ssl23_peek (analogous to ssl23_read), which previously
962 *) Replace rdtsc with _emit statements for VC++ version 5.
963 [Jeremy Cooper <jeremy@baymoo.org>]
965 +) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
966 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
969 +) Modify EVP_Digest*() routines so they now return values. Although the
970 internal software routines can never fail additional hardware versions
974 +) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
976 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
977 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
980 ERR_R_NESTED_ASN1_ERROR
982 ERR_R_MISSING_ASN1_EOS
983 were 4 .. 9, conflicting with
984 ERR_LIB_RSA (= ERR_R_RSA_LIB)
986 ERR_LIB_PEM (= ERR_R_PEM_LIB).
987 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
989 Add new error code 'ERR_R_INTERNAL_ERROR'.
992 +) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
996 +) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
997 sets the subject name for a new request or supersedes the
998 subject name in a given request. Formats that can be parsed are
999 'CN=Some Name, OU=myOU, C=IT'
1001 'CN=Some Name/OU=myOU/C=IT'.
1003 Add options '-batch' and '-verbose' to 'openssl req'.
1004 [Massimiliano Pala <madwolf@hackmasters.net>]
1006 *) Make it possible to reuse SSLv2 sessions.
1009 +) Introduce the possibility to access global variables through
1010 functions on platform were that's the best way to handle exporting
1011 global variables in shared libraries. To enable this functionality,
1012 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
1013 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
1014 is normally done by Configure or something similar).
1016 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
1017 in the source file (foo.c) like this:
1019 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
1020 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
1022 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
1023 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
1025 OPENSSL_DECLARE_GLOBAL(int,foo);
1026 #define foo OPENSSL_GLOBAL_REF(foo)
1027 OPENSSL_DECLARE_GLOBAL(double,bar);
1028 #define bar OPENSSL_GLOBAL_REF(bar)
1030 The #defines are very important, and therefore so is including the
1031 header file everywhere where the defined globals are used.
1033 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
1034 of ASN.1 items, but that structure is a bit different.
1036 The largest change is in util/mkdef.pl which has been enhanced with
1037 better and easier to understand logic to choose which symbols should
1038 go into the Windows .def files as well as a number of fixes and code
1039 cleanup (among others, algorithm keywords are now sorted
1040 lexicographically to avoid constant rewrites).
1043 *) In copy_email() check for >= 0 as a return value for
1044 X509_NAME_get_index_by_NID() since 0 is a valid index.
1045 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
1047 +) In BN_div() keep a copy of the sign of 'num' before writing the
1048 result to 'rm' because if rm==num the value will be overwritten
1049 and produce the wrong result if 'num' is negative: this caused
1050 problems with BN_mod() and BN_nnmod().
1053 +) Function OCSP_request_verify(). This checks the signature on an
1054 OCSP request and verifies the signer certificate. The signer
1055 certificate is just checked for a generic purpose and OCSP request
1059 +) Add OCSP_check_validity() function to check the validity of OCSP
1060 responses. OCSP responses are prepared in real time and may only
1061 be a few seconds old. Simply checking that the current time lies
1062 between thisUpdate and nextUpdate max reject otherwise valid responses
1063 caused by either OCSP responder or client clock inaccuracy. Instead
1064 we allow thisUpdate and nextUpdate to fall within a certain period of
1065 the current time. The age of the response can also optionally be
1066 checked. Two new options -validity_period and -status_age added to
1070 +) If signature or public key algorithm is unrecognized print out its
1071 OID rather that just UNKNOWN.
1074 *) Avoid coredump with unsupported or invalid public keys by checking if
1075 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
1076 PKCS7_verify() fails with non detached data.
1079 +) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
1080 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
1081 ID to be generated from the issuer certificate alone which can then be
1082 passed to OCSP_id_issuer_cmp().
1085 +) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
1086 ASN1 modules to export functions returning ASN1_ITEM pointers
1087 instead of the ASN1_ITEM structures themselves. This adds several
1088 new macros which allow the underlying ASN1 function/structure to
1089 be accessed transparently. As a result code should not use ASN1_ITEM
1090 references directly (such as &X509_it) but instead use the relevant
1091 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
1092 use of the new ASN1 code on platforms where exporting structures
1093 is problematical (for example in shared libraries) but exporting
1094 functions returning pointers to structures is not.
1097 +) Add support for overriding the generation of SSL/TLS session IDs.
1098 These callbacks can be registered either in an SSL_CTX or per SSL.
1099 The purpose of this is to allow applications to control, if they wish,
1100 the arbitrary values chosen for use as session IDs, particularly as it
1101 can be useful for session caching in multiple-server environments. A
1102 command-line switch for testing this (and any client code that wishes
1103 to use such a feature) has been added to "s_server".
1104 [Geoff Thorpe, Lutz Jaenicke]
1106 +) Modify mkdef.pl to recognise and parse preprocessor conditionals
1107 of the form '#if defined(...) || defined(...) || ...' and
1108 '#if !defined(...) && !defined(...) && ...'. This also avoids
1109 the growing number of special cases it was previously handling.
1112 *) Don't use getenv in library functions when run as setuid/setgid.
1113 New function OPENSSL_issetugid().
1116 +) Make all configuration macros available for application by making
1117 sure they are available in opensslconf.h, by giving them names starting
1118 with "OPENSSL_" to avoid conflicts with other packages and by making
1119 sure e_os2.h will cover all platform-specific cases together with
1121 Additionally, it is now possible to define configuration/platform-
1122 specific names (called "system identities"). In the C code, these
1123 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
1124 macro with the name beginning with "OPENSSL_SYS_", which is determined
1125 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
1129 +) New option -set_serial to 'req' and 'x509' this allows the serial
1130 number to use to be specified on the command line. Previously self
1131 signed certificates were hard coded with serial number 0 and the
1132 CA options of 'x509' had to use a serial number in a file which was
1136 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
1137 due to incorrect handling of multi-threading:
1139 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
1141 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
1143 3. Count how many times MemCheck_off() has been called so that
1144 nested use can be treated correctly. This also avoids
1145 inband-signalling in the previous code (which relied on the
1146 assumption that thread ID 0 is impossible).
1149 +) New options to 'ca' utility to support V2 CRL entry extensions.
1150 Currently CRL reason, invalidity date and hold instruction are
1151 supported. Add new CRL extensions to V3 code and some new objects.
1154 *) Add "-rand" option also to s_client and s_server.
1157 +) New function EVP_CIPHER_CTX_set_padding() this is used to
1158 disable standard block padding (aka PKCS#5 padding) in the EVP
1159 API, which was previously mandatory. This means that the data is
1160 not padded in any way and so the total length much be a multiple
1161 of the block size, otherwise an error occurs.
1164 +) Initial (incomplete) OCSP SSL support.
1167 *) Fix CPU detection on Irix 6.x.
1168 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
1169 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
1171 +) New function OCSP_parse_url(). This splits up a URL into its host,
1172 port and path components: primarily to parse OCSP URLs. New -url
1173 option to ocsp utility.
1176 +) New nonce behavior. The return value of OCSP_check_nonce() now
1177 reflects the various checks performed. Applications can decide
1178 whether to tolerate certain situations such as an absent nonce
1179 in a response when one was present in a request: the ocsp application
1180 just prints out a warning. New function OCSP_add1_basic_nonce()
1181 this is to allow responders to include a nonce in a response even if
1182 the request is nonce-less.
1185 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
1188 [This change does not apply to 0.9.7.]
1190 *) Use the cached encoding of an X509_NAME structure rather than
1191 copying it. This is apparently the reason for the libsafe "errors"
1192 but the code is actually correct.
1195 +) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
1196 skipped when using openssl x509 multiple times on a single input file,
1197 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
1200 +) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
1201 set string type: to handle setting ASN1_TIME structures. Fix ca
1202 utility to correctly initialize revocation date of CRLs.
1205 +) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
1206 the clients preferred ciphersuites and rather use its own preferences.
1207 Should help to work around M$ SGC (Server Gated Cryptography) bug in
1208 Internet Explorer by ensuring unchanged hash method during stepup.
1209 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
1212 +) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
1213 to aes and add a new 'exist' option to print out symbols that don't
1217 +) Additional options to ocsp utility to allow flags to be set and
1218 additional certificates supplied.
1221 +) Add the option -VAfile to 'openssl ocsp', so the user can give the
1222 OCSP client a number of certificate to only verify the response
1226 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
1227 Bleichenbacher's DSA attack.
1228 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
1229 to be set and top=0 forces the highest bit to be set; top=-1 is new
1230 and leaves the highest bit random.
1231 [Ulf Moeller, Bodo Moeller]
1233 +) Update Rijndael code to version 3.0 and change EVP AES ciphers to
1234 handle the new API. Currently only ECB, CBC modes supported. Add new
1235 AES OIDs. Add TLS AES ciphersuites as described in the "AES Ciphersuites
1236 for TLS" draft-ietf-tls-ciphersuite-03.txt.
1237 [Ben Laurie, Steve Henson]
1239 *) In the NCONF_...-based implementations for CONF_... queries
1240 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
1241 a temporary CONF structure with the data component set to NULL
1242 (which gives segmentation faults in lh_retrieve).
1243 Instead, use NULL for the CONF pointer in CONF_get_string and
1244 CONF_get_number (which may use environment variables) and directly
1245 return NULL from CONF_get_section.
1248 *) Fix potential buffer overrun for EBCDIC.
1251 +) New function OCSP_copy_nonce() to copy nonce value (if present) from
1252 request to response.
1255 +) Functions for OCSP responders. OCSP_request_onereq_count(),
1256 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
1257 extract information from a certificate request. OCSP_response_create()
1258 creates a response and optionally adds a basic response structure.
1259 OCSP_basic_add1_status() adds a complete single response to a basic
1260 response and returns the OCSP_SINGLERESP structure just added (to allow
1261 extensions to be included for example). OCSP_basic_add1_cert() adds a
1262 certificate to a basic response and OCSP_basic_sign() signs a basic
1263 response with various flags. New helper functions ASN1_TIME_check()
1264 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
1265 (converts ASN1_TIME to GeneralizedTime).
1268 +) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
1269 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
1270 structure from a certificate. X509_pubkey_digest() digests the public_key
1271 contents: this is used in various key identifiers.
1274 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
1275 keyUsage if basicConstraints absent for a CA.
1278 *) Make SMIME_write_PKCS7() write mail header values with a format that
1279 is more generally accepted (no spaces before the semicolon), since
1280 some programs can't parse those values properly otherwise. Also make
1281 sure BIO's that break lines after each write do not create invalid
1285 +) Make sk_sort() tolerate a NULL argument.
1286 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
1288 +) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
1289 passed by the function are trusted implicitly. If any of them signed the
1290 response then it is assumed to be valid and is not verified.
1293 -) Make the CRL encoding routines work with empty SEQUENCE OF. The
1294 macros previously used would not encode an empty SEQUENCE OF
1295 and break the signature.
1297 [This change does not apply to 0.9.7.]
1299 *) Zero the premaster secret after deriving the master secret in
1303 +) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
1304 to data. This was previously part of the PKCS7 ASN1 code. This
1305 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
1306 [Steve Henson, reported by Kenneth R. Robinette
1307 <support@securenetterm.com>]
1309 +) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
1310 routines: without these tracing memory leaks is very painful.
1311 Fix leaks in PKCS12 and PKCS7 routines.
1314 *) Add some EVP_add_digest_alias registrations (as found in
1315 OpenSSL_add_all_digests()) to SSL_library_init()
1316 aka OpenSSL_add_ssl_algorithms(). This provides improved
1317 compatibility with peers using X.509 certificates
1318 with unconventional AlgorithmIdentifier OIDs.
1321 *) Fix for Irix with NO_ASM.
1322 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
1324 *) ./config script fixes.
1325 [Ulf Moeller, Richard Levitte]
1327 +) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
1328 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
1329 effectively meant GeneralizedTime would never be used. Now it
1330 is initialised to -1 but X509_time_adj() now has to check the value
1331 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
1332 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
1333 [Steve Henson, reported by Kenneth R. Robinette
1334 <support@securenetterm.com>]
1336 +) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
1337 result in a zero length in the ASN1_INTEGER structure which was
1338 not consistent with the structure when d2i_ASN1_INTEGER() was used
1339 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
1340 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
1341 where it did not print out a minus for negative ASN1_INTEGER.
1344 *) Fix 'openssl passwd -1'.
1347 +) Add summary printout to ocsp utility. The various functions which
1348 convert status values to strings have been renamed to:
1349 OCSP_response_status_str(), OCSP_cert_status_str() and
1350 OCSP_crl_reason_str() and are no longer static. New options
1351 to verify nonce values and to disable verification. OCSP response
1352 printout format cleaned up.
1355 +) Add additional OCSP certificate checks. These are those specified
1356 in RFC2560. This consists of two separate checks: the CA of the
1357 certificate being checked must either be the OCSP signer certificate
1358 or the issuer of the OCSP signer certificate. In the latter case the
1359 OCSP signer certificate must contain the OCSP signing extended key
1360 usage. This check is performed by attempting to match the OCSP
1361 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
1362 in the OCSP_CERTID structures of the response.
1365 +) Initial OCSP certificate verification added to OCSP_basic_verify()
1366 and related routines. This uses the standard OpenSSL certificate
1367 verify routines to perform initial checks (just CA validity) and
1368 to obtain the certificate chain. Then additional checks will be
1369 performed on the chain. Currently the root CA is checked to see
1370 if it is explicitly trusted for OCSP signing. This is used to set
1371 a root CA as a global signing root: that is any certificate that
1372 chains to that CA is an acceptable OCSP signing certificate.
1375 +) New '-extfile ...' option to 'openssl ca' for reading X.509v3
1376 extensions from a separate configuration file.
1377 As when reading extensions from the main configuration file,
1378 the '-extensions ...' option may be used for specifying the
1380 [Massimiliano Pala <madwolf@comune.modena.it>]
1382 *) Change PKCS12_key_gen_asc() so it can cope with non null
1383 terminated strings whose length is passed in the passlen
1384 parameter, for example from PEM callbacks. This was done
1385 by adding an extra length parameter to asc2uni().
1386 [Steve Henson, reported by <oddissey@samsung.co.kr>]
1388 +) New OCSP utility. Allows OCSP requests to be generated or
1389 read. The request can be sent to a responder and the output
1390 parsed, outputed or printed in text form. Not complete yet:
1391 still needs to check the OCSP response validity.
1394 +) New subcommands for 'openssl ca':
1395 'openssl ca -status <serial>' prints the status of the cert with
1396 the given serial number (according to the index file).
1397 'openssl ca -updatedb' updates the expiry status of certificates
1399 [Massimiliano Pala <madwolf@comune.modena.it>]
1401 +) New '-newreq-nodes' command option to CA.pl. This is like
1402 '-newreq', but calls 'openssl req' with the '-nodes' option
1403 so that the resulting key is not encrypted.
1404 [Damien Miller <djm@mindrot.org>]
1406 +) New configuration for the GNU Hurd.
1407 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
1409 +) Initial code to implement OCSP basic response verify. This
1410 is currently incomplete. Currently just finds the signer's
1411 certificate and verifies the signature on the response.
1414 +) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
1415 value of OPENSSLDIR. This is available via the new '-d' option
1416 to 'openssl version', and is also included in 'openssl version -a'.
1419 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
1420 call failed, free the DSA structure.
1423 +) Allowing defining memory allocation callbacks that will be given
1424 file name and line number information in additional arguments
1425 (a const char* and an int). The basic functionality remains, as
1426 well as the original possibility to just replace malloc(),
1427 realloc() and free() by functions that do not know about these
1428 additional arguments. To register and find out the current
1429 settings for extended allocation functions, the following
1430 functions are provided:
1432 CRYPTO_set_mem_ex_functions
1433 CRYPTO_set_locked_mem_ex_functions
1434 CRYPTO_get_mem_ex_functions
1435 CRYPTO_get_locked_mem_ex_functions
1437 These work the same way as CRYPTO_set_mem_functions and friends.
1438 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
1439 extended allocation function is enabled.
1440 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
1441 a conventional allocation function is enabled.
1442 [Richard Levitte, Bodo Moeller]
1444 *) Fix to uni2asc() to cope with zero length Unicode strings.
1445 These are present in some PKCS#12 files.
1448 +) Finish off removing the remaining LHASH function pointer casts.
1449 There should no longer be any prototype-casting required when using
1450 the LHASH abstraction, and any casts that remain are "bugs". See
1451 the callback types and macros at the head of lhash.h for details
1452 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
1455 +) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
1456 If /dev/[u]random devices are not available or do not return enough
1457 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
1459 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
1460 /etc/entropy will be queried once each in this sequence, quering stops
1461 when enough entropy was collected without querying more sockets.
1464 +) Change the Unix RAND_poll() variant to be able to poll several
1465 random devices, as specified by DEVRANDOM, until a sufficient amount
1466 of data has been collected. We spend at most 10 ms on each file
1467 (select timeout) and read in non-blocking mode. DEVRANDOM now
1468 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
1469 (previously it was just the string "/dev/urandom"), so on typical
1470 platforms the 10 ms delay will never occur.
1471 Also separate out the Unix variant to its own file, rand_unix.c.
1472 For VMS, there's a currently-empty rand_vms.c.
1475 +) Move OCSP client related routines to ocsp_cl.c. These
1476 provide utility functions which an application needing
1477 to issue a request to an OCSP responder and analyse the
1478 response will typically need: as opposed to those which an
1479 OCSP responder itself would need which will be added later.
1481 OCSP_request_sign() signs an OCSP request with an API similar
1482 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
1483 response. OCSP_response_get1_basic() extracts basic response
1484 from response. OCSP_resp_find_status(): finds and extracts status
1485 information from an OCSP_CERTID structure (which will be created
1486 when the request structure is built). These are built from lower
1487 level functions which work on OCSP_SINGLERESP structures but
1488 wont normally be used unless the application wishes to examine
1489 extensions in the OCSP response for example.
1491 Replace nonce routines with a pair of functions.
1492 OCSP_request_add1_nonce() adds a nonce value and optionally
1493 generates a random value. OCSP_check_nonce() checks the
1494 validity of the nonce in an OCSP response.
1497 +) Change function OCSP_request_add() to OCSP_request_add0_id().
1498 This doesn't copy the supplied OCSP_CERTID and avoids the
1499 need to free up the newly created id. Change return type
1500 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
1501 This can then be used to add extensions to the request.
1502 Deleted OCSP_request_new(), since most of its functionality
1503 is now in OCSP_REQUEST_new() (and the case insensitive name
1504 clash) apart from the ability to set the request name which
1505 will be added elsewhere.
1508 +) Update OCSP API. Remove obsolete extensions argument from
1509 various functions. Extensions are now handled using the new
1510 OCSP extension code. New simple OCSP HTTP function which
1511 can be used to send requests and parse the response.
1514 +) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
1515 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
1516 uses the special reorder version of SET OF to sort the attributes
1517 and reorder them to match the encoded order. This resolves a long
1518 standing problem: a verify on a PKCS7 structure just after signing
1519 it used to fail because the attribute order did not match the
1520 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
1521 it uses the received order. This is necessary to tolerate some broken
1522 software that does not order SET OF. This is handled by encoding
1523 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
1524 to produce the required SET OF.
1527 +) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
1528 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
1529 files to get correct declarations of the ASN.1 item variables.
1532 +) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
1533 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
1534 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
1535 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
1536 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
1537 ASN1_ITEM and no wrapper functions.
1540 +) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
1541 replace the old function pointer based I/O routines. Change most of
1542 the *_d2i_bio() and *_d2i_fp() functions to use these.
1545 +) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
1546 lines, recognice more "algorithms" that can be deselected, and make
1547 it complain about algorithm deselection that isn't recognised.
1550 +) New ASN1 functions to handle dup, sign, verify, digest, pack and
1551 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
1552 to use new functions. Add NO_ASN1_OLD which can be set to remove
1553 some old style ASN1 functions: this can be used to determine if old
1554 code will still work when these eventually go away.
1557 +) New extension functions for OCSP structures, these follow the
1558 same conventions as certificates and CRLs.
1561 +) New function X509V3_add1_i2d(). This automatically encodes and
1562 adds an extension. Its behaviour can be customised with various
1563 flags to append, replace or delete. Various wrappers added for
1564 certifcates and CRLs.
1567 +) Fix to avoid calling the underlying ASN1 print routine when
1568 an extension cannot be parsed. Correct a typo in the
1569 OCSP_SERVICELOC extension. Tidy up print OCSP format.
1572 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
1573 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
1574 when writing a 32767 byte record.
1575 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
1577 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
1578 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
1580 (RSA objects have a reference count access to which is protected
1581 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
1582 so they are meant to be shared between threads.)
1583 [Bodo Moeller, Geoff Thorpe; original patch submitted by
1584 "Reddie, Steven" <Steven.Reddie@ca.com>]
1586 +) Make mkdef.pl parse some of the ASN1 macros and add apropriate
1587 entries for variables.
1590 *) Fix a deadlock in CRYPTO_mem_leaks().
1593 +) Add functionality to apps/openssl.c for detecting locking
1594 problems: As the program is single-threaded, all we have
1595 to do is register a locking callback using an array for
1596 storing which locks are currently held by the program.
1599 +) Use a lock around the call to CRYPTO_get_ex_new_index() in
1600 SSL_get_ex_data_X509_STORE_idx(), which is used in
1601 ssl_verify_cert_chain() and thus can be called at any time
1602 during TLS/SSL handshakes so that thread-safety is essential.
1603 Unfortunately, the ex_data design is not at all suited
1604 for multi-threaded use, so it probably should be abolished.
1607 +) Added Broadcom "ubsec" ENGINE to OpenSSL.
1608 [Broadcom, tweaked and integrated by Geoff Thorpe]
1610 +) Move common extension printing code to new function
1611 X509V3_print_extensions(). Reorganise OCSP print routines and
1612 implement some needed OCSP ASN1 functions. Add OCSP extensions.
1615 +) New function X509_signature_print() to remove duplication in some
1619 +) Add a special meaning when SET OF and SEQUENCE OF flags are both
1620 set (this was treated exactly the same as SET OF previously). This
1621 is used to reorder the STACK representing the structure to match the
1622 encoding. This will be used to get round a problem where a PKCS7
1623 structure which was signed could not be verified because the STACK
1624 order did not reflect the encoded order.
1627 +) Reimplement the OCSP ASN1 module using the new code.
1630 +) Update the X509V3 code to permit the use of an ASN1_ITEM structure
1631 for its ASN1 operations. The old style function pointers still exist
1632 for now but they will eventually go away.
1635 +) Merge in replacement ASN1 code from the ASN1 branch. This almost
1636 completely replaces the old ASN1 functionality with a table driven
1637 encoder and decoder which interprets an ASN1_ITEM structure describing
1638 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
1639 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
1640 has also been converted to the new form.
1643 +) Change BN_mod_exp_recp so that negative moduli are tolerated
1644 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
1645 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
1646 for negative moduli.
1649 +) Fix BN_uadd and BN_usub: Always return non-negative results instead
1650 of not touching the result's sign bit.
1653 +) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
1657 +) Changed the LHASH code to use prototypes for callbacks, and created
1658 macros to declare and implement thin (optionally static) functions
1659 that provide type-safety and avoid function pointer casting for the
1660 type-specific callbacks.
1663 *) Use better test patterns in bntest.
1666 +) Added Kerberos Cipher Suites to be used with TLS, as written in
1668 [Veers Staats <staatsvr@asc.hpc.mil>,
1669 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
1671 *) rand_win.c fix for Borland C.
1674 *) BN_rshift bugfix for n == 0.
1677 +) Reformat the FAQ so the different questions and answers can be divided
1678 in sections depending on the subject.
1681 +) Have the zlib compression code load ZLIB.DLL dynamically under
1685 +) New function BN_mod_sqrt for computing square roots modulo a prime
1686 (using the probabilistic Tonelli-Shanks algorithm unless
1687 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
1688 be handled deterministically).
1689 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
1691 *) Add a 'bctest' script that checks for some known 'bc' bugs
1692 so that 'make test' does not abort just because 'bc' is broken.
1695 *) Store verify_result within SSL_SESSION also for client side to
1696 avoid potential security hole. (Re-used sessions on the client side
1697 always resulted in verify_result==X509_V_OK, not using the original
1698 result of the server certificate verification.)
1701 +) Make BN_mod_inverse faster by explicitly handling small quotients
1702 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
1703 512 bits], about 30% for larger ones [1024 or 2048 bits].)
1706 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
1707 SSL3_RT_APPLICATION_DATA, return 0.
1708 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
1712 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
1713 releases, have been re-implemented by renaming the previous
1714 implementations of ssl2_read and ssl3_read to ssl2_read_internal
1715 and ssl3_read_internal, respectively, and adding 'peek' parameters
1716 to them. The new ssl[23]_{read,peek} functions are calls to
1717 ssl[23]_read_internal with the 'peek' flag set appropriately.
1718 A 'peek' parameter has also been added to ssl3_read_bytes, which
1719 does the actual work for ssl3_read_internal.
1722 +) New function BN_kronecker.
1725 +) Fix BN_gcd so that it works on negative inputs; the result is
1726 positive unless both parameters are zero.
1727 Previously something reasonably close to an infinite loop was
1728 possible because numbers could be growing instead of shrinking
1729 in the implementation of Euclid's algorithm.
1732 +) Fix BN_is_word() and BN_is_one() macros to take into account the
1733 sign of the number in question.
1735 Fix BN_is_word(a,w) to work correctly for w == 0.
1737 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
1738 because its test if the absolute value of 'a' equals 'w'.
1739 Note that BN_abs_is_word does *not* handle w == 0 reliably;
1740 it exists mostly for use in the implementations of BN_is_zero(),
1741 BN_is_one(), and BN_is_word().
1744 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
1745 the method-specific "init()" handler. Also clean up ex_data after
1746 calling the method-specific "finish()" handler. Previously, this was
1747 happening the other way round.
1750 +) New function BN_swap.
1753 +) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
1754 the exponentiation functions are more likely to produce reasonable
1755 results on negative inputs.
1758 +) Change BN_mod_mul so that the result is always non-negative.
1759 Previously, it could be negative if one of the factors was negative;
1760 I don't think anyone really wanted that behaviour.
1763 +) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
1764 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
1765 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
1766 and add new functions:
1775 BN_mod_lshift1_quick
1779 These functions always generate non-negative results.
1781 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
1782 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
1784 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
1785 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
1786 be reduced modulo m.
1787 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
1789 +) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
1790 was actually never needed) and in BN_mul(). The removal in BN_mul()
1791 required a small change in bn_mul_part_recursive() and the addition
1792 of the functions bn_cmp_part_words(), bn_sub_part_words() and
1793 bn_add_part_words(), which do the same thing as bn_cmp_words(),
1794 bn_sub_words() and bn_add_words() except they take arrays with
1798 +) In 'openssl passwd', verify passwords read from the terminal
1799 unless the '-salt' option is used (which usually means that
1800 verification would just waste user's time since the resulting
1801 hash is going to be compared with some given password hash)
1802 or the new '-noverify' option is used.
1804 This is an incompatible change, but it does not affect
1805 non-interactive use of 'openssl passwd' (passwords on the command
1806 line, '-stdin' option, '-in ...' option) and thus should not
1810 +) Remove all references to RSAref, since there's no more need for it.
1813 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
1814 The previous value, 12, was not always sufficient for BN_mod_exp().
1817 +) Make DSO load along a path given through an environment variable
1818 (SHLIB_PATH) with shl_load().
1821 +) Constify the ENGINE code as a result of BIGNUM constification.
1822 Also constify the RSA code and most things related to it. In a
1823 few places, most notable in the depth of the ASN.1 code, ugly
1824 casts back to non-const were required (to be solved at a later
1828 +) Make it so the openssl application has all engines loaded by default.
1831 +) Constify the BIGNUM routines a little more.
1834 *) Make sure that shared libraries get the internal name engine with
1835 the full version number and not just 0. This should mark the
1836 shared libraries as not backward compatible. Of course, this should
1837 be changed again when we can guarantee backward binary compatibility.
1840 +) Add the following functions:
1842 ENGINE_load_cswift()
1844 ENGINE_load_atalla()
1846 ENGINE_load_builtin_engines()
1848 That way, an application can itself choose if external engines that
1849 are built-in in OpenSSL shall ever be used or not. The benefit is
1850 that applications won't have to be linked with libdl or other dso
1851 libraries unless it's really needed.
1853 Changed 'openssl engine' to load all engines on demand.
1854 Changed the engine header files to avoid the duplication of some
1855 declarations (they differed!).
1858 +) 'openssl engine' can now list capabilities.
1861 +) Better error reporting in 'openssl engine'.
1864 +) Never call load_dh_param(NULL) in s_server.
1867 +) Add engine application. It can currently list engines by name and
1868 identity, and test if they are actually available.
1871 +) Add support for shared libraries under Irix.
1872 [Albert Chin-A-Young <china@thewrittenword.com>]
1874 +) Improve RPM specification file by forcing symbolic linking and making
1875 sure the installed documentation is also owned by root.root.
1876 [Damien Miller <djm@mindrot.org>]
1878 +) Add configuration option to build on Linux on both big-endian and
1880 [Ralf Baechle <ralf@uni-koblenz.de>]
1882 +) Give the OpenSSL applications more possibilities to make use of
1883 keys (public as well as private) handled by engines.
1886 +) Add OCSP code that comes from CertCo.
1889 +) Add VMS support for the Rijndael code.
1892 +) Added untested support for Nuron crypto accelerator.
1895 +) Add support for external cryptographic devices. This code was
1896 previously distributed separately as the "engine" branch.
1897 [Geoff Thorpe, Richard Levitte]
1899 +) Rework the filename-translation in the DSO code. It is now possible to
1900 have far greater control over how a "name" is turned into a filename
1901 depending on the operating environment and any oddities about the
1902 different shared library filenames on each system.
1905 +) Support threads on FreeBSD-elf in Configure.
1908 +) Add the possibility to create shared libraries on HP-UX
1911 +) Fix for SHA1 assembly problem with MASM: it produces
1912 warnings about corrupt line number information when assembling
1913 with debugging information. This is caused by the overlapping
1915 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
1918 NCONF_get_number() has no error checking at all. As a replacement,
1919 NCONF_get_number_e() is defined (_e for "error checking") and is
1920 promoted strongly. The old NCONF_get_number is kept around for
1921 binary backward compatibility.
1922 Make it possible for methods to load from something other than a BIO,
1923 by providing a function pointer that is given a name instead of a BIO.
1924 For example, this could be used to load configuration data from an
1928 *) Fix typo in get_cert_by_subject() in by_dir.c
1929 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
1931 *) Rework the system to generate shared libraries:
1933 - Make note of the expected extension for the shared libraries and
1934 if there is a need for symbolic links from for example libcrypto.so.0
1935 to libcrypto.so.0.9.7. There is extended info in Configure for
1938 - Make as few rebuilds of the shared libraries as possible.
1940 - Still avoid linking the OpenSSL programs with the shared libraries.
1942 - When installing, install the shared libraries separately from the
1946 +) Fix for non blocking accept BIOs. Added new I/O special reason
1947 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
1948 with non blocking I/O was not possible because no retry code was
1949 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
1953 +) Added the beginnings of Rijndael support.
1956 +) Fix for bug in DirectoryString mask setting. Add support for
1957 X509_NAME_print_ex() in 'req' and X509_print_ex() function
1958 to allow certificate printing to more controllable, additional
1959 'certopt' option to 'x509' to allow new printing options to be
1963 +) Clean old EAY MD5 hack from e_os.h.
1966 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
1968 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
1969 and not in SSL_clear because the latter is also used by the
1970 accept/connect functions; previously, the settings made by
1971 SSL_set_read_ahead would be lost during the handshake.
1972 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
1974 *) Correct util/mkdef.pl to be selective about disabled algorithms.
1975 Previously, it would create entries for disableed algorithms no
1979 *) Added several new manual pages for SSL_* function.
1982 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
1984 *) In ssl23_get_client_hello, generate an error message when faced
1985 with an initial SSL 3.0/TLS record that is too small to contain the
1986 first two bytes of the ClientHello message, i.e. client_version.
1987 (Note that this is a pathologic case that probably has never happened
1988 in real life.) The previous approach was to use the version number
1989 from the record header as a substitute; but our protocol choice
1990 should not depend on that one because it is not authenticated
1991 by the Finished messages.
1994 *) More robust randomness gathering functions for Windows.
1995 [Jeffrey Altman <jaltman@columbia.edu>]
1997 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
1998 not set then we don't setup the error code for issuer check errors
1999 to avoid possibly overwriting other errors which the callback does
2000 handle. If an application does set the flag then we assume it knows
2001 what it is doing and can handle the new informational codes
2005 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
2006 a general "ANY" type, as such it should be able to decode anything
2007 including tagged types. However it didn't check the class so it would
2008 wrongly interpret tagged types in the same way as their universal
2009 counterpart and unknown types were just rejected. Changed so that the
2010 tagged and unknown types are handled in the same way as a SEQUENCE:
2011 that is the encoding is stored intact. There is also a new type
2012 "V_ASN1_OTHER" which is used when the class is not universal, in this
2013 case we have no idea what the actual type is so we just lump them all
2017 *) On VMS, stdout may very well lead to a file that is written to
2018 in a record-oriented fashion. That means that every write() will
2019 write a separate record, which will be read separately by the
2020 programs trying to read from it. This can be very confusing.
2022 The solution is to put a BIO filter in the way that will buffer
2023 text until a linefeed is reached, and then write everything a
2024 line at a time, so every record written will be an actual line,
2025 not chunks of lines and not (usually doesn't happen, but I've
2026 seen it once) several lines in one record. BIO_f_linebuffer() is
2029 Currently, it's a VMS-only method, because that's where it has
2030 been tested well enough.
2033 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
2034 it can return incorrect results.
2035 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
2036 but it was in 0.9.6-beta[12].)
2039 *) Disable the check for content being present when verifying detached
2040 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
2041 include zero length content when signing messages.
2044 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
2045 BIO_ctrl (for BIO pairs).
2048 *) Add DSO method for VMS.
2051 *) Bug fix: Montgomery multiplication could produce results with the
2055 *) Add RPM specification openssl.spec and modify it to build three
2056 packages. The default package contains applications, application
2057 documentation and run-time libraries. The devel package contains
2058 include files, static libraries and function documentation. The
2059 doc package contains the contents of the doc directory. The original
2060 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
2063 *) Add a large number of documentation files for many SSL routines.
2064 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
2066 *) Add a configuration entry for Sony News 4.
2067 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
2069 *) Don't set the two most significant bits to one when generating a
2070 random number < q in the DSA library.
2073 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
2074 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
2075 the underlying transport is blocking) if a handshake took place.
2076 (The default behaviour is needed by applications such as s_client
2077 and s_server that use select() to determine when to use SSL_read;
2078 but for applications that know in advance when to expect data, it
2079 just makes things more complicated.)
2082 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
2086 *) Add a few more EBCDIC conditionals that make `req' and `x509'
2087 work better on such systems.
2088 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
2090 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
2091 Update PKCS12_parse() so it copies the friendlyName and the
2092 keyid to the certificates aux info.
2095 *) Fix bug in PKCS7_verify() which caused an infinite loop
2096 if there was more than one signature.
2097 [Sven Uszpelkat <su@celocom.de>]
2099 *) Major change in util/mkdef.pl to include extra information
2100 about each symbol, as well as presentig variables as well
2101 as functions. This change means that there's n more need
2102 to rebuild the .num files when some algorithms are excluded.
2105 *) Allow the verify time to be set by an application,
2106 rather than always using the current time.
2109 *) Phase 2 verify code reorganisation. The certificate
2110 verify code now looks up an issuer certificate by a
2111 number of criteria: subject name, authority key id
2112 and key usage. It also verifies self signed certificates
2113 by the same criteria. The main comparison function is
2114 X509_check_issued() which performs these checks.
2116 Lot of changes were necessary in order to support this
2117 without completely rewriting the lookup code.
2119 Authority and subject key identifier are now cached.
2121 The LHASH 'certs' is X509_STORE has now been replaced
2122 by a STACK_OF(X509_OBJECT). This is mainly because an
2123 LHASH can't store or retrieve multiple objects with
2124 the same hash value.
2126 As a result various functions (which were all internal
2127 use only) have changed to handle the new X509_STORE
2128 structure. This will break anything that messed round
2129 with X509_STORE internally.
2131 The functions X509_STORE_add_cert() now checks for an
2132 exact match, rather than just subject name.
2134 The X509_STORE API doesn't directly support the retrieval
2135 of multiple certificates matching a given criteria, however
2136 this can be worked round by performing a lookup first
2137 (which will fill the cache with candidate certificates)
2138 and then examining the cache for matches. This is probably
2139 the best we can do without throwing out X509_LOOKUP
2140 entirely (maybe later...).
2142 The X509_VERIFY_CTX structure has been enhanced considerably.
2144 All certificate lookup operations now go via a get_issuer()
2145 callback. Although this currently uses an X509_STORE it
2146 can be replaced by custom lookups. This is a simple way
2147 to bypass the X509_STORE hackery necessary to make this
2148 work and makes it possible to use more efficient techniques
2149 in future. A very simple version which uses a simple
2150 STACK for its trusted certificate store is also provided
2151 using X509_STORE_CTX_trusted_stack().
2153 The verify_cb() and verify() callbacks now have equivalents
2154 in the X509_STORE_CTX structure.
2156 X509_STORE_CTX also has a 'flags' field which can be used
2157 to customise the verify behaviour.
2160 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
2161 excludes S/MIME capabilities.
2164 *) When a certificate request is read in keep a copy of the
2165 original encoding of the signed data and use it when outputing
2166 again. Signatures then use the original encoding rather than
2167 a decoded, encoded version which may cause problems if the
2168 request is improperly encoded.
2171 *) For consistency with other BIO_puts implementations, call
2172 buffer_write(b, ...) directly in buffer_puts instead of calling
2175 In BIO_puts, increment b->num_write as in BIO_write.
2176 [Peter.Sylvester@EdelWeb.fr]
2178 *) Fix BN_mul_word for the case where the word is 0. (We have to use
2179 BN_zero, we may not return a BIGNUM with an array consisting of
2183 *) Avoid calling abort() from within the library when problems are
2184 detected, except if preprocessor symbols have been defined
2185 (such as REF_CHECK, BN_DEBUG etc.).
2188 *) New openssl application 'rsautl'. This utility can be
2189 used for low level RSA operations. DER public key
2190 BIO/fp routines also added.
2193 *) New Configure entry and patches for compiling on QNX 4.
2194 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
2196 *) A demo state-machine implementation was sponsored by
2197 Nuron (http://www.nuron.com/) and is now available in
2198 demos/state_machine.
2201 *) New options added to the 'dgst' utility for signature
2202 generation and verification.
2205 *) Unrecognized PKCS#7 content types are now handled via a
2206 catch all ASN1_TYPE structure. This allows unsupported
2207 types to be stored as a "blob" and an application can
2208 encode and decode it manually.
2211 *) Fix various signed/unsigned issues to make a_strex.c
2213 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
2215 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
2216 length if passed a buffer. ASN1_INTEGER_to_BN failed
2217 if passed a NULL BN and its argument was negative.
2218 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
2220 *) Modification to PKCS#7 encoding routines to output definite
2221 length encoding. Since currently the whole structures are in
2222 memory there's not real point in using indefinite length
2223 constructed encoding. However if OpenSSL is compiled with
2224 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
2227 *) Added BIO_vprintf() and BIO_vsnprintf().
2230 *) Added more prefixes to parse for in the the strings written
2231 through a logging bio, to cover all the levels that are available
2232 through syslog. The prefixes are now:
2234 PANIC, EMERG, EMR => LOG_EMERG
2235 ALERT, ALR => LOG_ALERT
2236 CRIT, CRI => LOG_CRIT
2237 ERROR, ERR => LOG_ERR
2238 WARNING, WARN, WAR => LOG_WARNING
2239 NOTICE, NOTE, NOT => LOG_NOTICE
2240 INFO, INF => LOG_INFO
2241 DEBUG, DBG => LOG_DEBUG
2243 and as before, if none of those prefixes are present at the
2244 beginning of the string, LOG_ERR is chosen.
2246 On Win32, the LOG_* levels are mapped according to this:
2248 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
2249 LOG_WARNING => EVENTLOG_WARNING_TYPE
2250 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
2254 *) Made it possible to reconfigure with just the configuration
2255 argument "reconf" or "reconfigure". The command line arguments
2256 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
2257 and are retrieved from there when reconfiguring.
2261 [Assar Westerlund <assar@sics.se>, Richard Levitte]
2263 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
2266 *) The obj_dat.pl script was messing up the sorting of object
2267 names. The reason was that it compared the quoted version
2268 of strings as a result "OCSP" > "OCSP Signing" because
2269 " > SPACE. Changed script to store unquoted versions of
2270 names and add quotes on output. It was also omitting some
2271 names from the lookup table if they were given a default
2272 value (that is if SN is missing it is given the same
2273 value as LN and vice versa), these are now added on the
2274 grounds that if an object has a name we should be able to
2275 look it up. Finally added warning output when duplicate
2276 short or long names are found.
2279 *) Changes needed for Tandem NSK.
2280 [Scott Uroff <scott@xypro.com>]
2282 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
2283 RSA_padding_check_SSLv23(), special padding was never detected
2284 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
2285 version rollback attacks was not effective.
2287 In s23_clnt.c, don't use special rollback-attack detection padding
2288 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
2289 client; similarly, in s23_srvr.c, don't do the rollback check if
2290 SSL 2.0 is the only protocol enabled in the server.
2293 *) Make it possible to get hexdumps of unprintable data with 'openssl
2294 asn1parse'. By implication, the functions ASN1_parse_dump() and
2295 BIO_dump_indent() are added.
2298 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
2299 these print out strings and name structures based on various
2300 flags including RFC2253 support and proper handling of
2301 multibyte characters. Added options to the 'x509' utility
2302 to allow the various flags to be set.
2305 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
2306 Also change the functions X509_cmp_current_time() and
2307 X509_gmtime_adj() work with an ASN1_TIME structure,
2308 this will enable certificates using GeneralizedTime in validity
2309 dates to be checked.
2312 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
2313 negative public key encodings) on by default,
2314 NO_NEG_PUBKEY_BUG can be set to disable it.
2317 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
2318 content octets. An i2c_ASN1_OBJECT is unnecessary because
2319 the encoding can be trivially obtained from the structure.
2322 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
2323 not read locks (CRYPTO_r_[un]lock).
2326 *) A first attempt at creating official support for shared
2327 libraries through configuration. I've kept it so the
2328 default is static libraries only, and the OpenSSL programs
2329 are always statically linked for now, but there are
2330 preparations for dynamic linking in place.
2331 This has been tested on Linux and Tru64.
2334 *) Randomness polling function for Win9x, as described in:
2335 Peter Gutmann, Software Generation of Practically Strong
2339 *) Fix so PRNG is seeded in req if using an already existing
2343 *) New options to smime application. -inform and -outform
2344 allow alternative formats for the S/MIME message including
2345 PEM and DER. The -content option allows the content to be
2346 specified separately. This should allow things like Netscape
2347 form signing output easier to verify.
2350 *) Fix the ASN1 encoding of tags using the 'long form'.
2353 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
2354 STRING types. These convert content octets to and from the
2355 underlying type. The actual tag and length octets are
2356 already assumed to have been read in and checked. These
2357 are needed because all other string types have virtually
2358 identical handling apart from the tag. By having versions
2359 of the ASN1 functions that just operate on content octets
2360 IMPLICIT tagging can be handled properly. It also allows
2361 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
2362 and ASN1_INTEGER are identical apart from the tag.
2365 *) Change the handling of OID objects as follows:
2367 - New object identifiers are inserted in objects.txt, following
2368 the syntax given in objects.README.
2369 - objects.pl is used to process obj_mac.num and create a new
2371 - obj_dat.pl is used to create a new obj_dat.h, using the data in
2374 This is currently kind of a hack, and the perl code in objects.pl
2375 isn't very elegant, but it works as I intended. The simplest way
2376 to check that it worked correctly is to look in obj_dat.h and
2377 check the array nid_objs and make sure the objects haven't moved
2378 around (this is important!). Additions are OK, as well as
2379 consistent name changes.
2382 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
2385 *) Addition of the command line parameter '-rand file' to 'openssl req'.
2386 The given file adds to whatever has already been seeded into the
2387 random pool through the RANDFILE configuration file option or
2388 environment variable, or the default random state file.
2391 *) mkstack.pl now sorts each macro group into lexical order.
2392 Previously the output order depended on the order the files
2393 appeared in the directory, resulting in needless rewriting
2397 *) Patches to make OpenSSL compile under Win32 again. Mostly
2398 work arounds for the VC++ problem that it treats func() as
2399 func(void). Also stripped out the parts of mkdef.pl that
2400 added extra typesafe functions: these no longer exist.
2403 *) Reorganisation of the stack code. The macros are now all
2404 collected in safestack.h . Each macro is defined in terms of
2405 a "stack macro" of the form SKM_<name>(type, a, b). The
2406 DEBUG_SAFESTACK is now handled in terms of function casts,
2407 this has the advantage of retaining type safety without the
2408 use of additional functions. If DEBUG_SAFESTACK is not defined
2409 then the non typesafe macros are used instead. Also modified the
2410 mkstack.pl script to handle the new form. Needs testing to see
2411 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
2412 the default if no major problems. Similar behaviour for ASN1_SET_OF
2413 and PKCS12_STACK_OF.
2416 *) When some versions of IIS use the 'NET' form of private key the
2417 key derivation algorithm is different. Normally MD5(password) is
2418 used as a 128 bit RC4 key. In the modified case
2419 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
2420 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
2421 as the old Netscape_RSA functions except they have an additional
2422 'sgckey' parameter which uses the modified algorithm. Also added
2423 an -sgckey command line option to the rsa utility. Thanks to
2424 Adrian Peck <bertie@ncipher.com> for posting details of the modified
2425 algorithm to openssl-dev.
2428 *) The evp_local.h macros were using 'c.##kname' which resulted in
2429 invalid expansion on some systems (SCO 5.0.5 for example).
2430 Corrected to 'c.kname'.
2431 [Phillip Porch <root@theporch.com>]
2433 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
2434 a STACK of email addresses from a certificate or request, these look
2435 in the subject name and the subject alternative name extensions and
2436 omit any duplicate addresses.
2439 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
2440 This makes DSA verification about 2 % faster.
2443 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
2444 (meaning that now 2^5 values will be precomputed, which is only 4 KB
2445 plus overhead for 1024 bit moduli).
2446 This makes exponentiations about 0.5 % faster for 1024 bit
2447 exponents (as measured by "openssl speed rsa2048").
2450 *) Rename memory handling macros to avoid conflicts with other
2452 Malloc => OPENSSL_malloc
2453 Malloc_locked => OPENSSL_malloc_locked
2454 Realloc => OPENSSL_realloc
2455 Free => OPENSSL_free
2458 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
2459 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
2462 *) CygWin32 support.
2463 [John Jarvie <jjarvie@newsguy.com>]
2465 *) The type-safe stack code has been rejigged. It is now only compiled
2466 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
2467 by default all type-specific stack functions are "#define"d back to
2468 standard stack functions. This results in more streamlined output
2469 but retains the type-safety checking possibilities of the original
2473 *) The STACK code has been cleaned up, and certain type declarations
2474 that didn't make a lot of sense have been brought in line. This has
2475 also involved a cleanup of sorts in safestack.h to more correctly
2476 map type-safe stack functions onto their plain stack counterparts.
2477 This work has also resulted in a variety of "const"ifications of
2478 lots of the code, especially "_cmp" operations which should normally
2479 be prototyped with "const" parameters anyway.
2482 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
2483 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
2484 (The PRNG state consists of two parts, the large pool 'state' and 'md',
2485 where all of 'md' is used each time the PRNG is used, but 'state'
2486 is used only indexed by a cyclic counter. As entropy may not be
2487 well distributed from the beginning, 'md' is important as a
2488 chaining variable. However, the output function chains only half
2489 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
2490 all of 'md', and seeding with STATE_SIZE dummy bytes will result
2491 in all of 'state' being rewritten, with the new values depending
2492 on virtually all of 'md'. This overcomes the 80 bit limitation.)
2495 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
2496 the handshake is continued after ssl_verify_cert_chain();
2497 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
2498 can lead to 'unexplainable' connection aborts later.
2499 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
2501 *) Major EVP API cipher revision.
2502 Add hooks for extra EVP features. This allows various cipher
2503 parameters to be set in the EVP interface. Support added for variable
2504 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
2505 setting of RC2 and RC5 parameters.
2507 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
2510 Remove lots of duplicated code from the EVP library. For example *every*
2511 cipher init() function handles the 'iv' in the same way according to the
2512 cipher mode. They also all do nothing if the 'key' parameter is NULL and
2513 for CFB and OFB modes they zero ctx->num.
2515 New functionality allows removal of S/MIME code RC2 hack.
2517 Most of the routines have the same form and so can be declared in terms
2520 By shifting this to the top level EVP_CipherInit() it can be removed from
2521 all individual ciphers. If the cipher wants to handle IVs or keys
2522 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
2525 Change lots of functions like EVP_EncryptUpdate() to now return a
2526 value: although software versions of the algorithms cannot fail
2527 any installed hardware versions can.
2530 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
2531 this option is set, tolerate broken clients that send the negotiated
2532 protocol version number instead of the requested protocol version
2536 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
2537 i.e. non-zero for export ciphersuites, zero otherwise.
2538 Previous versions had this flag inverted, inconsistent with
2539 rsa_tmp_cb (..._TMP_RSA_CB).
2540 [Bodo Moeller; problem reported by Amit Chopra]
2542 *) Add missing DSA library text string. Work around for some IIS
2543 key files with invalid SEQUENCE encoding.
2546 *) Add a document (doc/standards.txt) that list all kinds of standards
2547 and so on that are implemented in OpenSSL.
2550 *) Enhance c_rehash script. Old version would mishandle certificates
2551 with the same subject name hash and wouldn't handle CRLs at all.
2552 Added -fingerprint option to crl utility, to support new c_rehash
2556 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
2559 *) Fix for SSL server purpose checking. Server checking was
2560 rejecting certificates which had extended key usage present
2561 but no ssl client purpose.
2562 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
2564 *) Make PKCS#12 code work with no password. The PKCS#12 spec
2565 is a little unclear about how a blank password is handled.
2566 Since the password in encoded as a BMPString with terminating
2567 double NULL a zero length password would end up as just the
2568 double NULL. However no password at all is different and is
2569 handled differently in the PKCS#12 key generation code. NS
2570 treats a blank password as zero length. MSIE treats it as no
2571 password on export: but it will try both on import. We now do
2572 the same: PKCS12_parse() tries zero length and no password if
2573 the password is set to "" or NULL (NULL is now a valid password:
2574 it wasn't before) as does the pkcs12 application.
2577 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
2578 perror when PEM_read_bio_X509_REQ fails, the error message must
2579 be obtained from the error queue.
2582 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
2583 it in ERR_remove_state if appropriate, and change ERR_get_state
2584 accordingly to avoid race conditions (this is necessary because
2585 thread_hash is no longer constant once set).
2588 *) Bugfix for linux-elf makefile.one.
2591 *) RSA_get_default_method() will now cause a default
2592 RSA_METHOD to be chosen if one doesn't exist already.
2593 Previously this was only set during a call to RSA_new()
2594 or RSA_new_method(NULL) meaning it was possible for
2595 RSA_get_default_method() to return NULL.
2598 *) Added native name translation to the existing DSO code
2599 that will convert (if the flag to do so is set) filenames
2600 that are sufficiently small and have no path information
2601 into a canonical native form. Eg. "blah" converted to
2602 "libblah.so" or "blah.dll" etc.
2605 *) New function ERR_error_string_n(e, buf, len) which is like
2606 ERR_error_string(e, buf), but writes at most 'len' bytes
2607 including the 0 terminator. For ERR_error_string_n, 'buf'
2609 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
2611 *) CONF library reworked to become more general. A new CONF
2612 configuration file reader "class" is implemented as well as a
2613 new functions (NCONF_*, for "New CONF") to handle it. The now
2614 old CONF_* functions are still there, but are reimplemented to
2615 work in terms of the new functions. Also, a set of functions
2616 to handle the internal storage of the configuration data is
2617 provided to make it easier to write new configuration file
2618 reader "classes" (I can definitely see something reading a
2619 configuration file in XML format, for example), called _CONF_*,
2620 or "the configuration storage API"...
2622 The new configuration file reading functions are:
2624 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
2625 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
2627 NCONF_default, NCONF_WIN32
2629 NCONF_dump_fp, NCONF_dump_bio
2631 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
2632 NCONF_new creates a new CONF object. This works in the same way
2633 as other interfaces in OpenSSL, like the BIO interface.
2634 NCONF_dump_* dump the internal storage of the configuration file,
2635 which is useful for debugging. All other functions take the same
2636 arguments as the old CONF_* functions wth the exception of the
2637 first that must be a `CONF *' instead of a `LHASH *'.
2639 To make it easer to use the new classes with the old CONF_* functions,
2640 the function CONF_set_default_method is provided.
2643 *) Add '-tls1' option to 'openssl ciphers', which was already
2644 mentioned in the documentation but had not been implemented.
2645 (This option is not yet really useful because even the additional
2646 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
2649 *) Initial DSO code added into libcrypto for letting OpenSSL (and
2650 OpenSSL-based applications) load shared libraries and bind to
2651 them in a portable way.
2652 [Geoff Thorpe, with contributions from Richard Levitte]
2654 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
2656 *) Make sure _lrotl and _lrotr are only used with MSVC.
2658 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
2659 (the default implementation of RAND_status).
2661 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
2662 to '-clrext' (= clear extensions), as intended and documented.
2663 [Bodo Moeller; inconsistency pointed out by Michael Attili
2666 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
2667 was larger than the MD block size.
2668 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
2670 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
2671 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
2672 using the passed key: if the passed key was a private key the result
2673 of X509_print(), for example, would be to print out all the private key
2677 *) des_quad_cksum() byte order bug fix.
2678 [Ulf Möller, using the problem description in krb4-0.9.7, where
2679 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
2681 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
2683 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
2685 *) For easily testing in shell scripts whether some command
2686 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
2687 returns with exit code 0 iff no command of the given name is available.
2688 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
2689 the output goes to stdout and nothing is printed to stderr.
2690 Additional arguments are always ignored.
2692 Since for each cipher there is a command of the same name,
2693 the 'no-cipher' compilation switches can be tested this way.
2695 ('openssl no-XXX' is not able to detect pseudo-commands such
2696 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
2699 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
2702 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
2703 is set; it will be thrown away anyway because each handshake creates
2705 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
2706 to parameters -- in previous versions (since OpenSSL 0.9.3) the
2707 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
2708 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
2711 *) New s_client option -ign_eof: EOF at stdin is ignored, and
2712 'Q' and 'R' lose their special meanings (quit/renegotiate).
2713 This is part of what -quiet does; unlike -quiet, -ign_eof
2714 does not suppress any output.
2717 *) Add compatibility options to the purpose and trust code. The
2718 purpose X509_PURPOSE_ANY is "any purpose" which automatically
2719 accepts a certificate or CA, this was the previous behaviour,
2720 with all the associated security issues.
2722 X509_TRUST_COMPAT is the old trust behaviour: only and
2723 automatically trust self signed roots in certificate store. A
2724 new trust setting X509_TRUST_DEFAULT is used to specify that
2725 a purpose has no associated trust setting and it should instead
2726 use the value in the default purpose.
2729 *) Fix the PKCS#8 DSA private key code so it decodes keys again
2730 and fix a memory leak.
2733 *) In util/mkerr.pl (which implements 'make errors'), preserve
2734 reason strings from the previous version of the .c file, as
2735 the default to have only downcase letters (and digits) in
2736 automatically generated reasons codes is not always appropriate.
2739 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
2740 using strerror. Previously, ERR_reason_error_string() returned
2741 library names as reason strings for SYSerr; but SYSerr is a special
2742 case where small numbers are errno values, not library numbers.
2745 *) Add '-dsaparam' option to 'openssl dhparam' application. This
2746 converts DSA parameters into DH parameters. (When creating parameters,
2747 DSA_generate_parameters is used.)
2750 *) Include 'length' (recommended exponent length) in C code generated
2751 by 'openssl dhparam -C'.
2754 *) The second argument to set_label in perlasm was already being used
2755 so couldn't be used as a "file scope" flag. Moved to third argument
2759 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
2760 instead of RAND_bytes for encryption IVs and salts.
2763 *) Include RAND_status() into RAND_METHOD instead of implementing
2764 it only for md_rand.c Otherwise replacing the PRNG by calling
2765 RAND_set_rand_method would be impossible.
2768 *) Don't let DSA_generate_key() enter an infinite loop if the random
2769 number generation fails.
2772 *) New 'rand' application for creating pseudo-random output.
2775 *) Added configuration support for Linux/IA64
2776 [Rolf Haberrecker <rolf@suse.de>]
2778 *) Assembler module support for Mingw32.
2781 *) Shared library support for HPUX (in shlib/).
2782 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
2784 *) Shared library support for Solaris gcc.
2785 [Lutz Behnke <behnke@trustcenter.de>]
2787 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
2789 *) PKCS7_encrypt() was adding text MIME headers twice because they
2790 were added manually and by SMIME_crlf_copy().
2793 *) In bntest.c don't call BN_rand with zero bits argument.
2794 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
2796 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
2797 case was implemented. This caused BN_div_recp() to fail occasionally.
2800 *) Add an optional second argument to the set_label() in the perl
2801 assembly language builder. If this argument exists and is set
2802 to 1 it signals that the assembler should use a symbol whose
2803 scope is the entire file, not just the current function. This
2804 is needed with MASM which uses the format label:: for this scope.
2805 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
2807 *) Change the ASN1 types so they are typedefs by default. Before
2808 almost all types were #define'd to ASN1_STRING which was causing
2809 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
2813 *) Change names of new functions to the new get1/get0 naming
2814 convention: After 'get1', the caller owns a reference count
2815 and has to call ..._free; 'get0' returns a pointer to some
2816 data structure without incrementing reference counters.
2817 (Some of the existing 'get' functions increment a reference
2818 counter, some don't.)
2819 Similarly, 'set1' and 'add1' functions increase reference
2820 counters or duplicate objects.
2823 *) Allow for the possibility of temp RSA key generation failure:
2824 the code used to assume it always worked and crashed on failure.
2827 *) Fix potential buffer overrun problem in BIO_printf().
2828 [Ulf Möller, using public domain code by Patrick Powell; problem
2829 pointed out by David Sacerdote <das33@cornell.edu>]
2831 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
2832 RAND_egd() and RAND_status(). In the command line application,
2833 the EGD socket can be specified like a seed file using RANDFILE
2837 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
2838 Some CAs (e.g. Verisign) distribute certificates in this form.
2841 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
2842 list to exclude them. This means that no special compilation option
2843 is needed to use anonymous DH: it just needs to be included in the
2847 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
2848 EVP_MD_type. The old functionality is available in a new macro called
2849 EVP_MD_md(). Change code that uses it and update docs.
2852 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
2853 where the 'void *' argument is replaced by a function pointer argument.
2854 Previously 'void *' was abused to point to functions, which works on
2855 many platforms, but is not correct. As these functions are usually
2856 called by macros defined in OpenSSL header files, most source code
2857 should work without changes.
2860 *) <openssl/opensslconf.h> (which is created by Configure) now contains
2861 sections with information on -D... compiler switches used for
2862 compiling the library so that applications can see them. To enable
2863 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
2864 must be defined. E.g.,
2865 #define OPENSSL_ALGORITHM_DEFINES
2866 #include <openssl/opensslconf.h>
2867 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
2868 [Richard Levitte, Ulf and Bodo Möller]
2870 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
2874 *) Change the 'other' type in certificate aux info to a STACK_OF
2875 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
2876 the required ASN1 format: arbitrary types determined by an OID.
2879 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
2880 argument to 'req'. This is not because the function is newer or
2881 better than others it just uses the work 'NEW' in the certificate
2882 request header lines. Some software needs this.
2885 *) Reorganise password command line arguments: now passwords can be
2886 obtained from various sources. Delete the PEM_cb function and make
2887 it the default behaviour: i.e. if the callback is NULL and the
2888 usrdata argument is not NULL interpret it as a null terminated pass
2889 phrase. If usrdata and the callback are NULL then the pass phrase
2890 is prompted for as usual.
2893 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
2894 the support is automatically enabled. The resulting binaries will
2895 autodetect the card and use it if present.
2896 [Ben Laurie and Compaq Inc.]
2898 *) Work around for Netscape hang bug. This sends certificate request
2899 and server done in one record. Since this is perfectly legal in the
2900 SSL/TLS protocol it isn't a "bug" option and is on by default. See
2901 the bugs/SSLv3 entry for more info.
2904 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
2907 *) Add -rand argument to smime and pkcs12 applications and read/write
2911 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
2914 *) Add command line password options to the remaining applications.
2917 *) Bug fix for BN_div_recp() for numerators with an even number of
2921 *) More tests in bntest.c, and changed test_bn output.
2924 *) ./config recognizes MacOS X now.
2927 *) Bug fix for BN_div() when the first words of num and divsor are
2928 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
2931 *) Add support for various broken PKCS#8 formats, and command line
2932 options to produce them.
2935 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
2936 get temporary BIGNUMs from a BN_CTX.
2939 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
2943 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
2944 include a #define from the old name to the new. The original intent
2945 was that statically linked binaries could for example just call
2946 SSLeay_add_all_ciphers() to just add ciphers to the table and not
2947 link with digests. This never worked becayse SSLeay_add_all_digests()
2948 and SSLeay_add_all_ciphers() were in the same source file so calling
2949 one would link with the other. They are now in separate source files.
2952 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
2955 *) Use a less unusual form of the Miller-Rabin primality test (it used
2956 a binary algorithm for exponentiation integrated into the Miller-Rabin
2957 loop, our standard modexp algorithms are faster).
2960 *) Support for the EBCDIC character set completed.
2961 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
2963 *) Source code cleanups: use const where appropriate, eliminate casts,
2964 use void * instead of char * in lhash.
2967 *) Bugfix: ssl3_send_server_key_exchange was not restartable
2968 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
2969 this the server could overwrite ephemeral keys that the client
2973 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
2974 using 50 iterations of the Rabin-Miller test.
2976 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
2977 iterations of the Rabin-Miller test as required by the appendix
2978 to FIPS PUB 186[-1]) instead of DSA_is_prime.
2979 As BN_is_prime_fasttest includes trial division, DSA parameter
2980 generation becomes much faster.
2982 This implies a change for the callback functions in DSA_is_prime
2983 and DSA_generate_parameters: The callback function is called once
2984 for each positive witness in the Rabin-Miller test, not just
2985 occasionally in the inner loop; and the parameters to the
2986 callback function now provide an iteration count for the outer
2987 loop rather than for the current invocation of the inner loop.
2988 DSA_generate_parameters additionally can call the callback
2989 function with an 'iteration count' of -1, meaning that a
2990 candidate has passed the trial division test (when q is generated
2991 from an application-provided seed, trial division is skipped).
2994 *) New function BN_is_prime_fasttest that optionally does trial
2995 division before starting the Rabin-Miller test and has
2996 an additional BN_CTX * argument (whereas BN_is_prime always
2997 has to allocate at least one BN_CTX).
2998 'callback(1, -1, cb_arg)' is called when a number has passed the
2999 trial division stage.
3002 *) Fix for bug in CRL encoding. The validity dates weren't being handled
3006 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
3009 *) New function BN_pseudo_rand().
3012 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
3013 bignum version of BN_from_montgomery() with the working code from
3014 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
3018 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
3019 made it impossible to use the same SSL_SESSION data structure in
3020 SSL2 clients in multiple threads.
3023 *) The return value of RAND_load_file() no longer counts bytes obtained
3024 by stat(). RAND_load_file(..., -1) is new and uses the complete file
3025 to seed the PRNG (previously an explicit byte count was required).
3026 [Ulf Möller, Bodo Möller]
3028 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
3029 used (char *) instead of (void *) and had casts all over the place.
3032 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
3035 *) Retain source code compatibility for BN_prime_checks macro:
3036 BN_is_prime(..., BN_prime_checks, ...) now uses
3037 BN_prime_checks_for_size to determine the appropriate number of
3038 Rabin-Miller iterations.
3041 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
3042 DH_CHECK_P_NOT_SAFE_PRIME.
3043 (Check if this is true? OpenPGP calls them "strong".)
3046 *) Merge the functionality of "dh" and "gendh" programs into a new program
3047 "dhparam". The old programs are retained for now but will handle DH keys
3048 (instead of parameters) in future.
3051 *) Make the ciphers, s_server and s_client programs check the return values
3052 when a new cipher list is set.
3055 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
3056 ciphers. Before when the 56bit ciphers were enabled the sorting was
3059 The syntax for the cipher sorting has been extended to support sorting by
3060 cipher-strength (using the strength_bits hard coded in the tables).
3061 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
3063 Fix a bug in the cipher-command parser: when supplying a cipher command
3064 string with an "undefined" symbol (neither command nor alphanumeric
3065 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
3066 an error is flagged.
3068 Due to the strength-sorting extension, the code of the
3069 ssl_create_cipher_list() function was completely rearranged. I hope that
3070 the readability was also increased :-)
3071 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
3073 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
3074 for the first serial number and places 2 in the serial number file. This
3075 avoids problems when the root CA is created with serial number zero and
3076 the first user certificate has the same issuer name and serial number
3080 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
3081 the new code. Add documentation for this stuff.
3084 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
3085 X509_*() to X509at_*() on the grounds that they don't handle X509
3086 structures and behave in an analagous way to the X509v3 functions:
3087 they shouldn't be called directly but wrapper functions should be used
3090 So we also now have some wrapper functions that call the X509at functions
3091 when passed certificate requests. (TO DO: similar things can be done with
3092 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
3093 things. Some of these need some d2i or i2d and print functionality
3094 because they handle more complex structures.)
3097 *) Add missing #ifndefs that caused missing symbols when building libssl
3098 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
3099 NO_RSA in ssl/s2*.c.
3100 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
3102 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
3103 has a return value which indicates the quality of the random data
3104 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
3105 error queue. New function RAND_pseudo_bytes() generates output that is
3106 guaranteed to be unique but not unpredictable. RAND_add is like
3107 RAND_seed, but takes an extra argument for an entropy estimate
3108 (RAND_seed always assumes full entropy).
3111 *) Do more iterations of Rabin-Miller probable prime test (specifically,
3112 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
3113 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
3114 in crypto/bn/bn_prime.c for the complete table). This guarantees a
3115 false-positive rate of at most 2^-80 for random input.
3118 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
3121 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
3122 in the 0.9.5 release), this returns the chain
3123 from an X509_CTX structure with a dup of the stack and all
3124 the X509 reference counts upped: so the stack will exist
3125 after X509_CTX_cleanup() has been called. Modify pkcs12.c
3128 Also make SSL_SESSION_print() print out the verify return
3132 *) Add manpage for the pkcs12 command. Also change the default
3133 behaviour so MAC iteration counts are used unless the new
3134 -nomaciter option is used. This improves file security and
3135 only older versions of MSIE (4.0 for example) need it.
3138 *) Honor the no-xxx Configure options when creating .DEF files.
3141 *) Add PKCS#10 attributes to field table: challengePassword,
3142 unstructuredName and unstructuredAddress. These are taken from
3143 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
3144 international characters are used.
3146 More changes to X509_ATTRIBUTE code: allow the setting of types
3147 based on strings. Remove the 'loc' parameter when adding
3148 attributes because these will be a SET OF encoding which is sorted
3152 *) Initial changes to the 'req' utility to allow request generation
3153 automation. This will allow an application to just generate a template
3154 file containing all the field values and have req construct the
3157 Initial support for X509_ATTRIBUTE handling. Stacks of these are
3158 used all over the place including certificate requests and PKCS#7
3159 structures. They are currently handled manually where necessary with
3160 some primitive wrappers for PKCS#7. The new functions behave in a
3161 manner analogous to the X509 extension functions: they allow
3162 attributes to be looked up by NID and added.
3164 Later something similar to the X509V3 code would be desirable to
3165 automatically handle the encoding, decoding and printing of the
3166 more complex types. The string types like challengePassword can
3167 be handled by the string table functions.
3169 Also modified the multi byte string table handling. Now there is
3170 a 'global mask' which masks out certain types. The table itself
3171 can use the flag STABLE_NO_MASK to ignore the mask setting: this
3172 is useful when for example there is only one permissible type
3173 (as in countryName) and using the mask might result in no valid
3177 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
3178 SSL_get_peer_finished to allow applications to obtain the latest
3179 Finished messages sent to the peer or expected from the peer,
3180 respectively. (SSL_get_peer_finished is usually the Finished message
3181 actually received from the peer, otherwise the protocol will be aborted.)
3183 As the Finished message are message digests of the complete handshake
3184 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
3185 be used for external authentication procedures when the authentication
3186 provided by SSL/TLS is not desired or is not enough.
3189 *) Enhanced support for Alpha Linux is added. Now ./config checks if
3190 the host supports BWX extension and if Compaq C is present on the
3191 $PATH. Just exploiting of the BWX extension results in 20-30%
3192 performance kick for some algorithms, e.g. DES and RC4 to mention
3193 a couple. Compaq C in turn generates ~20% faster code for MD5 and
3197 *) Add support for MS "fast SGC". This is arguably a violation of the
3198 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
3199 weak crypto and after checking the certificate is SGC a second one
3200 with strong crypto. MS SGC stops the first handshake after receiving
3201 the server certificate message and sends a second client hello. Since
3202 a server will typically do all the time consuming operations before
3203 expecting any further messages from the client (server key exchange
3204 is the most expensive) there is little difference between the two.
3206 To get OpenSSL to support MS SGC we have to permit a second client
3207 hello message after we have sent server done. In addition we have to
3208 reset the MAC if we do get this second client hello.
3211 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
3212 if a DER encoded private key is RSA or DSA traditional format. Changed
3213 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
3214 format DER encoded private key. Newer code should use PKCS#8 format which
3215 has the key type encoded in the ASN1 structure. Added DER private key
3216 support to pkcs8 application.
3219 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
3220 ciphersuites has been selected (as required by the SSL 3/TLS 1
3221 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
3222 is set, we interpret this as a request to violate the specification
3223 (the worst that can happen is a handshake failure, and 'correct'
3224 behaviour would result in a handshake failure anyway).
3227 *) In SSL_CTX_add_session, take into account that there might be multiple
3228 SSL_SESSION structures with the same session ID (e.g. when two threads
3229 concurrently obtain them from an external cache).
3230 The internal cache can handle only one SSL_SESSION with a given ID,
3231 so if there's a conflict, we now throw out the old one to achieve
3235 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
3236 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
3237 some routines that use cipher OIDs: some ciphers do not have OIDs
3238 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
3242 *) Simplify the trust setting structure and code. Now we just have
3243 two sequences of OIDs for trusted and rejected settings. These will
3244 typically have values the same as the extended key usage extension
3245 and any application specific purposes.
3247 The trust checking code now has a default behaviour: it will just
3248 check for an object with the same NID as the passed id. Functions can
3249 be provided to override either the default behaviour or the behaviour
3250 for a given id. SSL client, server and email already have functions
3251 in place for compatibility: they check the NID and also return "trusted"
3252 if the certificate is self signed.
3255 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
3256 traditional format into an EVP_PKEY structure.
3259 *) Add a password callback function PEM_cb() which either prompts for
3260 a password if usr_data is NULL or otherwise assumes it is a null
3261 terminated password. Allow passwords to be passed on command line
3262 environment or config files in a few more utilities.
3265 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
3266 keys. Add some short names for PKCS#8 PBE algorithms and allow them
3267 to be specified on the command line for the pkcs8 and pkcs12 utilities.
3268 Update documentation.
3271 *) Support for ASN1 "NULL" type. This could be handled before by using
3272 ASN1_TYPE but there wasn't any function that would try to read a NULL
3273 and produce an error if it couldn't. For compatibility we also have
3274 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
3275 don't allocate anything because they don't need to.
3278 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
3280 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
3282 *) Rebuild of the memory allocation routines used by OpenSSL code and
3283 possibly others as well. The purpose is to make an interface that
3284 provide hooks so anyone can build a separate set of allocation and
3285 deallocation routines to be used by OpenSSL, for example memory
3286 pool implementations, or something else, which was previously hard
3287 since Malloc(), Realloc() and Free() were defined as macros having
3288 the values malloc, realloc and free, respectively (except for Win32
3289 compilations). The same is provided for memory debugging code.
3290 OpenSSL already comes with functionality to find memory leaks, but
3291 this gives people a chance to debug other memory problems.
3293 With these changes, a new set of functions and macros have appeared:
3295 CRYPTO_set_mem_debug_functions() [F]
3296 CRYPTO_get_mem_debug_functions() [F]
3297 CRYPTO_dbg_set_options() [F]
3298 CRYPTO_dbg_get_options() [F]
3299 CRYPTO_malloc_debug_init() [M]
3301 The memory debug functions are NULL by default, unless the library
3302 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
3303 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
3304 gives the standard debugging functions that come with OpenSSL) or
3305 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
3306 provided by the library user) must be used. When the standard
3307 debugging functions are used, CRYPTO_dbg_set_options can be used to
3308 request additional information:
3309 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
3310 the CRYPTO_MDEBUG_xxx macro when compiling the library.
3312 Also, things like CRYPTO_set_mem_functions will always give the
3313 expected result (the new set of functions is used for allocation
3314 and deallocation) at all times, regardless of platform and compiler
3317 To finish it up, some functions that were never use in any other
3318 way than through macros have a new API and new semantic:
3321 CRYPTO_dbg_realloc()
3324 All macros of value have retained their old syntax.
3325 [Richard Levitte and Bodo Moeller]
3327 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
3328 ordering of SMIMECapabilities wasn't in "strength order" and there
3329 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
3333 *) Some ASN1 types with illegal zero length encoding (INTEGER,
3334 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
3335 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
3337 *) Merge in my S/MIME library for OpenSSL. This provides a simple
3338 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
3339 functionality to handle multipart/signed properly) and a utility
3340 called 'smime' to call all this stuff. This is based on code I
3341 originally wrote for Celo who have kindly allowed it to be
3342 included in OpenSSL.
3345 *) Add variants des_set_key_checked and des_set_key_unchecked of
3346 des_set_key (aka des_key_sched). Global variable des_check_key
3347 decides which of these is called by des_set_key; this way
3348 des_check_key behaves as it always did, but applications and
3349 the library itself, which was buggy for des_check_key == 1,
3350 have a cleaner way to pick the version they need.
3353 *) New function PKCS12_newpass() which changes the password of a
3357 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
3358 dynamic mix. In both cases the ids can be used as an index into the
3359 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
3360 functions so they accept a list of the field values and the
3361 application doesn't need to directly manipulate the X509_TRUST
3365 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
3369 *) Modify the way the V3 extension code looks up extensions. This now
3370 works in a similar way to the object code: we have some "standard"
3371 extensions in a static table which is searched with OBJ_bsearch()
3372 and the application can add dynamic ones if needed. The file
3373 crypto/x509v3/ext_dat.h now has the info: this file needs to be
3374 updated whenever a new extension is added to the core code and kept
3375 in ext_nid order. There is a simple program 'tabtest.c' which checks
3376 this. New extensions are not added too often so this file can readily
3377 be maintained manually.
3379 There are two big advantages in doing things this way. The extensions
3380 can be looked up immediately and no longer need to be "added" using
3381 X509V3_add_standard_extensions(): this function now does nothing.
3382 [Side note: I get *lots* of email saying the extension code doesn't
3383 work because people forget to call this function]
3384 Also no dynamic allocation is done unless new extensions are added:
3385 so if we don't add custom extensions there is no need to call
3386 X509V3_EXT_cleanup().
3389 *) Modify enc utility's salting as follows: make salting the default. Add a
3390 magic header, so unsalted files fail gracefully instead of just decrypting
3391 to garbage. This is because not salting is a big security hole, so people
3392 should be discouraged from doing it.
3395 *) Fixes and enhancements to the 'x509' utility. It allowed a message
3396 digest to be passed on the command line but it only used this
3397 parameter when signing a certificate. Modified so all relevant
3398 operations are affected by the digest parameter including the
3399 -fingerprint and -x509toreq options. Also -x509toreq choked if a
3400 DSA key was used because it didn't fix the digest.
3403 *) Initial certificate chain verify code. Currently tests the untrusted
3404 certificates for consistency with the verify purpose (which is set
3405 when the X509_STORE_CTX structure is set up) and checks the pathlength.
3407 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
3408 this is because it will reject chains with invalid extensions whereas
3409 every previous version of OpenSSL and SSLeay made no checks at all.
3411 Trust code: checks the root CA for the relevant trust settings. Trust
3412 settings have an initial value consistent with the verify purpose: e.g.
3413 if the verify purpose is for SSL client use it expects the CA to be
3414 trusted for SSL client use. However the default value can be changed to
3415 permit custom trust settings: one example of this would be to only trust
3416 certificates from a specific "secure" set of CAs.
3418 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
3419 which should be used for version portability: especially since the
3420 verify structure is likely to change more often now.
3422 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
3423 to set them. If not set then assume SSL clients will verify SSL servers
3426 Two new options to the verify program: -untrusted allows a set of
3427 untrusted certificates to be passed in and -purpose which sets the
3428 intended purpose of the certificate. If a purpose is set then the
3429 new chain verify code is used to check extension consistency.
3432 *) Support for the authority information access extension.
3435 *) Modify RSA and DSA PEM read routines to transparently handle
3436 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
3437 public keys in a format compatible with certificate
3438 SubjectPublicKeyInfo structures. Unfortunately there were already
3439 functions called *_PublicKey_* which used various odd formats so
3440 these are retained for compatibility: however the DSA variants were
3441 never in a public release so they have been deleted. Changed dsa/rsa
3442 utilities to handle the new format: note no releases ever handled public
3443 keys so we should be OK.
3445 The primary motivation for this change is to avoid the same fiasco
3446 that dogs private keys: there are several incompatible private key
3447 formats some of which are standard and some OpenSSL specific and
3448 require various evil hacks to allow partial transparent handling and
3449 even then it doesn't work with DER formats. Given the option anything
3450 other than PKCS#8 should be dumped: but the other formats have to
3451 stay in the name of compatibility.
3453 With public keys and the benefit of hindsight one standard format
3454 is used which works with EVP_PKEY, RSA or DSA structures: though
3455 it clearly returns an error if you try to read the wrong kind of key.
3457 Added a -pubkey option to the 'x509' utility to output the public key.
3458 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
3459 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
3460 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
3461 that do the same as the EVP_PKEY_assign_*() except they up the
3462 reference count of the added key (they don't "swallow" the
3466 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
3467 CRLs would fail if the file contained no certificates or no CRLs:
3468 added a new function to read in both types and return the number
3469 read: this means that if none are read it will be an error. The
3470 DER versions of the certificate and CRL reader would always fail
3471 because it isn't possible to mix certificates and CRLs in DER format
3472 without choking one or the other routine. Changed this to just read
3473 a certificate: this is the best we can do. Also modified the code
3474 in apps/verify.c to take notice of return codes: it was previously
3475 attempting to read in certificates from NULL pointers and ignoring
3476 any errors: this is one reason why the cert and CRL reader seemed
3477 to work. It doesn't check return codes from the default certificate
3478 routines: these may well fail if the certificates aren't installed.
3481 *) Code to support otherName option in GeneralName.
3484 *) First update to verify code. Change the verify utility
3485 so it warns if it is passed a self signed certificate:
3486 for consistency with the normal behaviour. X509_verify
3487 has been modified to it will now verify a self signed
3488 certificate if *exactly* the same certificate appears
3489 in the store: it was previously impossible to trust a
3490 single self signed certificate. This means that:
3491 openssl verify ss.pem
3492 now gives a warning about a self signed certificate but
3493 openssl verify -CAfile ss.pem ss.pem
3497 *) For servers, store verify_result in SSL_SESSION data structure
3498 (and add it to external session representation).
3499 This is needed when client certificate verifications fails,
3500 but an application-provided verification callback (set by
3501 SSL_CTX_set_cert_verify_callback) allows accepting the session
3502 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
3503 but returns 1): When the session is reused, we have to set
3504 ssl->verify_result to the appropriate error code to avoid
3506 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
3508 *) Fix a bug in the new PKCS#7 code: it didn't consider the
3509 case in PKCS7_dataInit() where the signed PKCS7 structure
3510 didn't contain any existing data because it was being created.
3511 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
3513 *) Add a salt to the key derivation routines in enc.c. This
3514 forms the first 8 bytes of the encrypted file. Also add a
3515 -S option to allow a salt to be input on the command line.
3518 *) New function X509_cmp(). Oddly enough there wasn't a function
3519 to compare two certificates. We do this by working out the SHA1
3520 hash and comparing that. X509_cmp() will be needed by the trust
3524 *) SSL_get1_session() is like SSL_get_session(), but increments
3525 the reference count in the SSL_SESSION returned.
3526 [Geoff Thorpe <geoff@eu.c2.net>]
3528 *) Fix for 'req': it was adding a null to request attributes.
3529 Also change the X509_LOOKUP and X509_INFO code to handle
3530 certificate auxiliary information.
3533 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
3537 *) Add the possibility to add extra information to the memory leak
3538 detecting output, to form tracebacks, showing from where each
3539 allocation was originated: CRYPTO_push_info("constant string") adds
3540 the string plus current file name and line number to a per-thread
3541 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
3542 is like calling CYRPTO_pop_info() until the stack is empty.
3543 Also updated memory leak detection code to be multi-thread-safe.
3546 *) Add options -text and -noout to pkcs7 utility and delete the
3547 encryption options which never did anything. Update docs.
3550 *) Add options to some of the utilities to allow the pass phrase
3551 to be included on either the command line (not recommended on
3552 OSes like Unix) or read from the environment. Update the
3553 manpages and fix a few bugs.
3556 *) Add a few manpages for some of the openssl commands.
3559 *) Fix the -revoke option in ca. It was freeing up memory twice,
3560 leaking and not finding already revoked certificates.
3563 *) Extensive changes to support certificate auxiliary information.
3564 This involves the use of X509_CERT_AUX structure and X509_AUX
3565 functions. An X509_AUX function such as PEM_read_X509_AUX()
3566 can still read in a certificate file in the usual way but it
3567 will also read in any additional "auxiliary information". By
3568 doing things this way a fair degree of compatibility can be
3569 retained: existing certificates can have this information added
3570 using the new 'x509' options.
3572 Current auxiliary information includes an "alias" and some trust
3573 settings. The trust settings will ultimately be used in enhanced
3574 certificate chain verification routines: currently a certificate
3575 can only be trusted if it is self signed and then it is trusted
3579 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
3580 The problem was that one of the replacement routines had not been working
3581 since SSLeay releases. For now the offending routine has been replaced
3582 with non-optimised assembler. Even so, this now gives around 95%
3583 performance improvement for 1024 bit RSA signs.
3586 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
3587 handling. Most clients have the effective key size in bits equal to
3588 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
3589 A few however don't do this and instead use the size of the decrypted key
3590 to determine the RC2 key length and the AlgorithmIdentifier to determine
3591 the effective key length. In this case the effective key length can still
3592 be 40 bits but the key length can be 168 bits for example. This is fixed
3593 by manually forcing an RC2 key into the EVP_PKEY structure because the
3594 EVP code can't currently handle unusual RC2 key sizes: it always assumes
3595 the key length and effective key length are equal.
3598 *) Add a bunch of functions that should simplify the creation of
3599 X509_NAME structures. Now you should be able to do:
3600 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
3601 and have it automatically work out the correct field type and fill in
3602 the structures. The more adventurous can try:
3603 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
3604 and it will (hopefully) work out the correct multibyte encoding.
3607 *) Change the 'req' utility to use the new field handling and multibyte
3608 copy routines. Before the DN field creation was handled in an ad hoc
3609 way in req, ca, and x509 which was rather broken and didn't support
3610 BMPStrings or UTF8Strings. Since some software doesn't implement
3611 BMPStrings or UTF8Strings yet, they can be enabled using the config file
3612 using the dirstring_type option. See the new comment in the default
3613 openssl.cnf for more info.
3616 *) Make crypto/rand/md_rand.c more robust:
3617 - Assure unique random numbers after fork().
3618 - Make sure that concurrent threads access the global counter and
3619 md serializably so that we never lose entropy in them
3620 or use exactly the same state in multiple threads.
3621 Access to the large state is not always serializable because
3622 the additional locking could be a performance killer, and
3623 md should be large enough anyway.
3626 *) New file apps/app_rand.c with commonly needed functionality
3627 for handling the random seed file.
3629 Use the random seed file in some applications that previously did not:
3631 dsaparam -genkey (which also ignored its '-rand' option),
3634 x509 (when signing).
3635 Except on systems with /dev/urandom, it is crucial to have a random
3636 seed file at least for key creation, DSA signing, and for DH exchanges;
3637 for RSA signatures we could do without one.
3639 gendh and gendsa (unlike genrsa) used to read only the first byte
3640 of each file listed in the '-rand' option. The function as previously
3641 found in genrsa is now in app_rand.c and is used by all programs
3642 that support '-rand'.
3645 *) In RAND_write_file, use mode 0600 for creating files;
3646 don't just chmod when it may be too late.
3649 *) Report an error from X509_STORE_load_locations
3650 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
3653 *) New function ASN1_mbstring_copy() this copies a string in either
3654 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
3655 into an ASN1_STRING type. A mask of permissible types is passed
3656 and it chooses the "minimal" type to use or an error if not type
3660 *) Add function equivalents to the various macros in asn1.h. The old
3661 macros are retained with an M_ prefix. Code inside the library can
3662 use the M_ macros. External code (including the openssl utility)
3663 should *NOT* in order to be "shared library friendly".
3666 *) Add various functions that can check a certificate's extensions
3667 to see if it usable for various purposes such as SSL client,
3668 server or S/MIME and CAs of these types. This is currently
3669 VERY EXPERIMENTAL but will ultimately be used for certificate chain
3670 verification. Also added a -purpose flag to x509 utility to
3671 print out all the purposes.
3674 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
3678 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
3679 for, obtain and decode and extension and obtain its critical flag.
3680 This allows all the necessary extension code to be handled in a
3681 single function call.
3684 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
3685 platforms. See crypto/rc4/rc4_enc.c for further details.
3688 *) New -noout option to asn1parse. This causes no output to be produced
3689 its main use is when combined with -strparse and -out to extract data
3690 from a file (which may not be in ASN.1 format).
3693 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
3694 when producing the local key id.
3695 [Richard Levitte <levitte@stacken.kth.se>]
3697 *) New option -dhparam in s_server. This allows a DH parameter file to be
3698 stated explicitly. If it is not stated then it tries the first server
3699 certificate file. The previous behaviour hard coded the filename
3703 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
3704 a public key to be input or output. For example:
3705 openssl rsa -in key.pem -pubout -out pubkey.pem
3706 Also added necessary DSA public key functions to handle this.
3709 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
3710 in the message. This was handled by allowing
3711 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
3712 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
3714 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
3715 to the end of the strings whereas this didn't. This would cause problems
3716 if strings read with d2i_ASN1_bytes() were later modified.
3717 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
3719 *) Fix for base64 decode bug. When a base64 bio reads only one line of
3720 data and it contains EOF it will end up returning an error. This is
3721 caused by input 46 bytes long. The cause is due to the way base64
3722 BIOs find the start of base64 encoded data. They do this by trying a
3723 trial decode on each line until they find one that works. When they
3724 do a flag is set and it starts again knowing it can pass all the
3725 data directly through the decoder. Unfortunately it doesn't reset
3726 the context it uses. This means that if EOF is reached an attempt
3727 is made to pass two EOFs through the context and this causes the
3728 resulting error. This can also cause other problems as well. As is
3729 usual with these problems it takes *ages* to find and the fix is
3730 trivial: move one line.
3731 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
3733 *) Ugly workaround to get s_client and s_server working under Windows. The
3734 old code wouldn't work because it needed to select() on sockets and the
3735 tty (for keypresses and to see if data could be written). Win32 only
3736 supports select() on sockets so we select() with a 1s timeout on the
3737 sockets and then see if any characters are waiting to be read, if none
3738 are present then we retry, we also assume we can always write data to
3739 the tty. This isn't nice because the code then blocks until we've
3740 received a complete line of data and it is effectively polling the
3741 keyboard at 1s intervals: however it's quite a bit better than not
3742 working at all :-) A dedicated Windows application might handle this
3743 with an event loop for example.
3746 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
3747 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
3748 will be called when RSA_sign() and RSA_verify() are used. This is useful
3749 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
3750 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
3751 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
3752 This necessitated the support of an extra signature type NID_md5_sha1
3753 for SSL signatures and modifications to the SSL library to use it instead
3754 of calling RSA_public_decrypt() and RSA_private_encrypt().
3757 *) Add new -verify -CAfile and -CApath options to the crl program, these
3758 will lookup a CRL issuers certificate and verify the signature in a
3759 similar way to the verify program. Tidy up the crl program so it
3760 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
3761 less strict. It will now permit CRL extensions even if it is not
3762 a V2 CRL: this will allow it to tolerate some broken CRLs.
3765 *) Initialize all non-automatic variables each time one of the openssl
3766 sub-programs is started (this is necessary as they may be started
3767 multiple times from the "OpenSSL>" prompt).
3768 [Lennart Bang, Bodo Moeller]
3770 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
3771 removing all other RSA functionality (this is what NO_RSA does). This
3772 is so (for example) those in the US can disable those operations covered
3773 by the RSA patent while allowing storage and parsing of RSA keys and RSA
3777 *) Non-copying interface to BIO pairs.
3778 (still largely untested)
3781 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
3782 ASCII string. This was handled independently in various places before.
3785 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
3786 UTF8 strings a character at a time.
3789 *) Use client_version from client hello to select the protocol
3790 (s23_srvr.c) and for RSA client key exchange verification
3791 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
3794 *) Add various utility functions to handle SPKACs, these were previously
3795 handled by poking round in the structure internals. Added new function
3796 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
3797 print, verify and generate SPKACs. Based on an original idea from
3798 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
3801 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
3804 *) Allow the config file extension section to be overwritten on the
3805 command line. Based on an original idea from Massimiliano Pala
3806 <madwolf@comune.modena.it>. The new option is called -extensions
3807 and can be applied to ca, req and x509. Also -reqexts to override
3808 the request extensions in req and -crlexts to override the crl extensions
3812 *) Add new feature to the SPKAC handling in ca. Now you can include
3813 the same field multiple times by preceding it by "XXXX." for example:
3816 this is the same syntax as used in the req config file.
3819 *) Allow certificate extensions to be added to certificate requests. These
3820 are specified in a 'req_extensions' option of the req section of the
3821 config file. They can be printed out with the -text option to req but
3822 are otherwise ignored at present.
3825 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
3826 data read consists of only the final block it would not decrypted because
3827 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
3828 A misplaced 'break' also meant the decrypted final block might not be
3829 copied until the next read.
3832 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
3833 a few extra parameters to the DH structure: these will be useful if
3834 for example we want the value of 'q' or implement X9.42 DH.
3837 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
3838 provides hooks that allow the default DSA functions or functions on a
3839 "per key" basis to be replaced. This allows hardware acceleration and
3840 hardware key storage to be handled without major modification to the
3841 library. Also added low level modexp hooks and CRYPTO_EX structure and
3842 associated functions.
3845 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
3846 as "read only": it can't be written to and the buffer it points to will
3847 not be freed. Reading from a read only BIO is much more efficient than
3848 a normal memory BIO. This was added because there are several times when
3849 an area of memory needs to be read from a BIO. The previous method was
3850 to create a memory BIO and write the data to it, this results in two
3851 copies of the data and an O(n^2) reading algorithm. There is a new
3852 function BIO_new_mem_buf() which creates a read only memory BIO from
3853 an area of memory. Also modified the PKCS#7 routines to use read only
3857 *) Bugfix: ssl23_get_client_hello did not work properly when called in
3858 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
3859 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
3860 but a retry condition occured while trying to read the rest.
3863 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
3864 NID_pkcs7_encrypted by default: this was wrong since this should almost
3865 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
3866 the encrypted data type: this is a more sensible place to put it and it
3867 allows the PKCS#12 code to be tidied up that duplicated this
3871 *) Changed obj_dat.pl script so it takes its input and output files on
3872 the command line. This should avoid shell escape redirection problems
3876 *) Initial support for certificate extension requests, these are included
3877 in things like Xenroll certificate requests. Included functions to allow
3878 extensions to be obtained and added.
3881 *) -crlf option to s_client and s_server for sending newlines as
3882 CRLF (as required by many protocols).
3885 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
3887 *) Install libRSAglue.a when OpenSSL is built with RSAref.
3888 [Ralf S. Engelschall]
3890 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
3891 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
3893 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
3897 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
3898 DH parameters/keys (q is lost during that conversion, but the resulting
3899 DH parameters contain its length).
3901 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
3902 much faster than DH_generate_parameters (which creates parameters
3903 where p = 2*q + 1), and also the smaller q makes DH computations
3904 much more efficient (160-bit exponentiation instead of 1024-bit
3905 exponentiation); so this provides a convenient way to support DHE
3906 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
3907 utter importance to use
3908 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
3910 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
3911 when such DH parameters are used, because otherwise small subgroup
3912 attacks may become possible!
3915 *) Avoid memory leak in i2d_DHparams.
3918 *) Allow the -k option to be used more than once in the enc program:
3919 this allows the same encrypted message to be read by multiple recipients.
3922 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
3923 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
3924 it will always use the numerical form of the OID, even if it has a short
3928 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
3929 method only got called if p,q,dmp1,dmq1,iqmp components were present,
3930 otherwise bn_mod_exp was called. In the case of hardware keys for example
3931 no private key components need be present and it might store extra data
3932 in the RSA structure, which cannot be accessed from bn_mod_exp.
3933 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
3934 private key operations.
3937 *) Added support for SPARC Linux.
3940 *) pem_password_cb function type incompatibly changed from
3941 typedef int pem_password_cb(char *buf, int size, int rwflag);
3943 ....(char *buf, int size, int rwflag, void *userdata);
3944 so that applications can pass data to their callbacks:
3945 The PEM[_ASN1]_{read,write}... functions and macros now take an
3946 additional void * argument, which is just handed through whenever
3947 the password callback is called.
3948 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
3950 New function SSL_CTX_set_default_passwd_cb_userdata.
3952 Compatibility note: As many C implementations push function arguments
3953 onto the stack in reverse order, the new library version is likely to
3954 interoperate with programs that have been compiled with the old
3955 pem_password_cb definition (PEM_whatever takes some data that
3956 happens to be on the stack as its last argument, and the callback
3957 just ignores this garbage); but there is no guarantee whatsoever that
3960 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
3961 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
3962 problems not only on Windows, but also on some Unix platforms.
3963 To avoid problematic command lines, these definitions are now in an
3964 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
3965 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
3968 *) MIPS III/IV assembler module is reimplemented.
3971 *) More DES library cleanups: remove references to srand/rand and
3972 delete an unused file.
3975 *) Add support for the the free Netwide assembler (NASM) under Win32,
3976 since not many people have MASM (ml) and it can be hard to obtain.
3977 This is currently experimental but it seems to work OK and pass all
3978 the tests. Check out INSTALL.W32 for info.
3981 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
3982 without temporary keys kept an extra copy of the server key,
3983 and connections with temporary keys did not free everything in case
3987 *) New function RSA_check_key and new openssl rsa option -check
3988 for verifying the consistency of RSA keys.
3989 [Ulf Moeller, Bodo Moeller]
3991 *) Various changes to make Win32 compile work:
3992 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
3993 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
3994 comparison" warnings.
3995 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
3998 *) Add a debugging option to PKCS#5 v2 key generation function: when
3999 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
4000 derived keys are printed to stderr.
4003 *) Copy the flags in ASN1_STRING_dup().
4004 [Roman E. Pavlov <pre@mo.msk.ru>]
4006 *) The x509 application mishandled signing requests containing DSA
4007 keys when the signing key was also DSA and the parameters didn't match.
4009 It was supposed to omit the parameters when they matched the signing key:
4010 the verifying software was then supposed to automatically use the CA's
4011 parameters if they were absent from the end user certificate.
4013 Omitting parameters is no longer recommended. The test was also
4014 the wrong way round! This was probably due to unusual behaviour in
4015 EVP_cmp_parameters() which returns 1 if the parameters match.
4016 This meant that parameters were omitted when they *didn't* match and
4017 the certificate was useless. Certificates signed with 'ca' didn't have
4019 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
4021 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
4022 The interface is as follows:
4023 Applications can use
4024 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
4025 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
4026 "off" is now the default.
4027 The library internally uses
4028 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
4029 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
4030 to disable memory-checking temporarily.
4032 Some inconsistent states that previously were possible (and were
4033 even the default) are now avoided.
4035 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
4036 with each memory chunk allocated; this is occasionally more helpful
4037 than just having a counter.
4039 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
4041 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
4045 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
4046 which largely parallels "options", but is for changing API behaviour,
4047 whereas "options" are about protocol behaviour.
4048 Initial "mode" flags are:
4050 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
4051 a single record has been written.
4052 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
4053 retries use the same buffer location.
4054 (But all of the contents must be
4058 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
4061 *) Fix problems with no-hmac etc.
4062 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
4064 *) New functions RSA_get_default_method(), RSA_set_method() and
4065 RSA_get_method(). These allows replacement of RSA_METHODs without having
4066 to mess around with the internals of an RSA structure.
4069 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
4070 Also really enable memory leak checks in openssl.c and in some
4072 [Chad C. Mulligan, Bodo Moeller]
4074 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
4075 up the length of negative integers. This has now been simplified to just
4076 store the length when it is first determined and use it later, rather
4077 than trying to keep track of where data is copied and updating it to
4079 [Steve Henson, reported by Brien Wheeler
4080 <bwheeler@authentica-security.com>]
4082 *) Add a new function PKCS7_signatureVerify. This allows the verification
4083 of a PKCS#7 signature but with the signing certificate passed to the
4084 function itself. This contrasts with PKCS7_dataVerify which assumes the
4085 certificate is present in the PKCS#7 structure. This isn't always the
4086 case: certificates can be omitted from a PKCS#7 structure and be
4087 distributed by "out of band" means (such as a certificate database).
4090 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
4091 function prototypes in pem.h, also change util/mkdef.pl to add the
4092 necessary function names.
4095 *) mk1mf.pl (used by Windows builds) did not properly read the
4096 options set by Configure in the top level Makefile, and Configure
4097 was not even able to write more than one option correctly.
4098 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
4101 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
4102 file to be loaded from a BIO or FILE pointer. The BIO version will
4103 for example allow memory BIOs to contain config info.
4106 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
4107 Whoever hopes to achieve shared-library compatibility across versions
4108 must use this, not the compile-time macro.
4109 (Exercise 0.9.4: Which is the minimum library version required by
4111 Note: All this applies only to multi-threaded programs, others don't
4115 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
4116 through a BIO pair triggered the default case, i.e.
4117 SSLerr(...,SSL_R_UNKNOWN_STATE).
4120 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
4121 can use the SSL library even if none of the specific BIOs is
4125 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
4126 for the encoded length.
4127 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
4129 *) Add initial documentation of the X509V3 functions.
4132 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
4133 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
4134 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
4135 secure PKCS#8 private key format with a high iteration count.
4138 *) Fix determination of Perl interpreter: A perl or perl5
4139 _directory_ in $PATH was also accepted as the interpreter.
4140 [Ralf S. Engelschall]
4142 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
4143 wrong with it but it was very old and did things like calling
4144 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
4148 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
4149 to use the new extension code.
4152 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
4153 with macros. This should make it easier to change their form, add extra
4154 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
4158 *) Add to configuration table a new entry that can specify an alternative
4159 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
4160 according to Mark Crispin <MRC@Panda.COM>.
4164 *) DES CBC did not update the IV. Weird.
4167 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
4168 Changing the behaviour of the former might break existing programs --
4169 where IV updating is needed, des_ncbc_encrypt can be used.
4172 *) When bntest is run from "make test" it drives bc to check its
4173 calculations, as well as internally checking them. If an internal check
4174 fails, it needs to cause bc to give a non-zero result or make test carries
4175 on without noticing the failure. Fixed.
4178 *) DES library cleanups.
4181 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
4182 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
4183 ciphers. NOTE: although the key derivation function has been verified
4184 against some published test vectors it has not been extensively tested
4185 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
4189 *) Instead of "mkdir -p", which is not fully portable, use new
4190 Perl script "util/mkdir-p.pl".
4193 *) Rewrite the way password based encryption (PBE) is handled. It used to
4194 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
4195 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
4196 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
4197 the 'parameter' field of the AlgorithmIdentifier is passed to the
4198 underlying key generation function so it must do its own ASN1 parsing.
4199 This has also changed the EVP_PBE_CipherInit() function which now has a
4200 'parameter' argument instead of literal salt and iteration count values
4201 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
4204 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
4205 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
4206 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
4207 KEY" because this clashed with PKCS#8 unencrypted string. Since this
4208 value was just used as a "magic string" and not used directly its
4209 value doesn't matter.
4212 *) Introduce some semblance of const correctness to BN. Shame C doesn't
4216 *) "linux-sparc64" configuration (ultrapenguin).
4217 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
4218 "linux-sparc" configuration.
4219 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
4221 *) config now generates no-xxx options for missing ciphers.
4224 *) Support the EBCDIC character set (work in progress).
4225 File ebcdic.c not yet included because it has a different license.
4226 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
4228 *) Support BS2000/OSD-POSIX.
4229 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
4231 *) Make callbacks for key generation use void * instead of char *.
4234 *) Make S/MIME samples compile (not yet tested).
4237 *) Additional typesafe stacks.
4240 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
4244 Changes between 0.9.3 and 0.9.3a [29 May 1999]
4246 *) New configuration variant "sco5-gcc".
4248 *) Updated some demos.
4249 [Sean O Riordain, Wade Scholine]
4251 *) Add missing BIO_free at exit of pkcs12 application.
4254 *) Fix memory leak in conf.c.
4257 *) Updates for Win32 to assembler version of MD5.
4260 *) Set #! path to perl in apps/der_chop to where we found it
4261 instead of using a fixed path.
4264 *) SHA library changes for irix64-mips4-cc.
4267 *) Improvements for VMS support.
4271 Changes between 0.9.2b and 0.9.3 [24 May 1999]
4273 *) Bignum library bug fix. IRIX 6 passes "make test" now!
4274 This also avoids the problems with SC4.2 and unpatched SC5.
4275 [Andy Polyakov <appro@fy.chalmers.se>]
4277 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
4278 These are required because of the typesafe stack would otherwise break
4279 existing code. If old code used a structure member which used to be STACK
4280 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
4281 sk_num or sk_value it would produce an error because the num, data members
4282 are not present in STACK_OF. Now it just produces a warning. sk_set
4283 replaces the old method of assigning a value to sk_value
4284 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
4285 that does this will no longer work (and should use sk_set instead) but
4286 this could be regarded as a "questionable" behaviour anyway.
4289 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
4290 correctly handle encrypted S/MIME data.
4293 *) Change type of various DES function arguments from des_cblock
4294 (which means, in function argument declarations, pointer to char)
4295 to des_cblock * (meaning pointer to array with 8 char elements),
4296 which allows the compiler to do more typechecking; it was like
4297 that back in SSLeay, but with lots of ugly casts.
4299 Introduce new type const_des_cblock.
4302 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
4303 problems: find RecipientInfo structure that matches recipient certificate
4304 and initialise the ASN1 structures properly based on passed cipher.
4307 *) Belatedly make the BN tests actually check the results.
4310 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
4311 to and from BNs: it was completely broken. New compilation option
4312 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
4313 key elements as negative integers.
4316 *) Reorganize and speed up MD5.
4317 [Andy Polyakov <appro@fy.chalmers.se>]
4320 [Richard Levitte <richard@levitte.org>]
4322 *) New option -out to asn1parse to allow the parsed structure to be
4323 output to a file. This is most useful when combined with the -strparse
4324 option to examine the output of things like OCTET STRINGS.
4327 *) Make SSL library a little more fool-proof by not requiring any longer
4328 that SSL_set_{accept,connect}_state be called before
4329 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
4330 in many applications because usually everything *appeared* to work as
4331 intended anyway -- now it really works as intended).
4334 *) Move openssl.cnf out of lib/.
4337 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
4338 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
4339 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
4340 [Ralf S. Engelschall]
4342 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
4343 handle PKCS#7 enveloped data properly.
4344 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
4346 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
4347 copying pointers. The cert_st handling is changed by this in
4348 various ways (and thus what used to be known as ctx->default_cert
4349 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
4350 any longer when s->cert does not give us what we need).
4351 ssl_cert_instantiate becomes obsolete by this change.
4352 As soon as we've got the new code right (possibly it already is?),
4353 we have solved a couple of bugs of the earlier code where s->cert
4354 was used as if it could not have been shared with other SSL structures.
4356 Note that using the SSL API in certain dirty ways now will result
4357 in different behaviour than observed with earlier library versions:
4358 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
4359 does not influence s as it used to.
4361 In order to clean up things more thoroughly, inside SSL_SESSION
4362 we don't use CERT any longer, but a new structure SESS_CERT
4363 that holds per-session data (if available); currently, this is
4364 the peer's certificate chain and, for clients, the server's certificate
4365 and temporary key. CERT holds only those values that can have
4366 meaningful defaults in an SSL_CTX.
4369 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
4370 from the internal representation. Various PKCS#7 fixes: remove some
4371 evil casts and set the enc_dig_alg field properly based on the signing
4375 *) Allow PKCS#12 password to be set from the command line or the
4376 environment. Let 'ca' get its config file name from the environment
4377 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
4381 *) Allow certificate policies extension to use an IA5STRING for the
4382 organization field. This is contrary to the PKIX definition but
4383 VeriSign uses it and IE5 only recognises this form. Document 'x509'
4387 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
4388 without disallowing inline assembler and the like for non-pedantic builds.
4391 *) Support Borland C++ builder.
4392 [Janez Jere <jj@void.si>, modified by Ulf Möller]
4397 *) SHA-1 cleanups and performance enhancements.
4398 [Andy Polyakov <appro@fy.chalmers.se>]
4400 *) Sparc v8plus assembler for the bignum library.
4401 [Andy Polyakov <appro@fy.chalmers.se>]
4403 *) Accept any -xxx and +xxx compiler options in Configure.
4406 *) Update HPUX configuration.
4409 *) Add missing sk_<type>_unshift() function to safestack.h
4410 [Ralf S. Engelschall]
4412 *) New function SSL_CTX_use_certificate_chain_file that sets the
4413 "extra_cert"s in addition to the certificate. (This makes sense
4414 only for "PEM" format files, as chains as a whole are not
4418 *) Support verify_depth from the SSL API.
4419 x509_vfy.c had what can be considered an off-by-one-error:
4420 Its depth (which was not part of the external interface)
4421 was actually counting the number of certificates in a chain;
4422 now it really counts the depth.
4425 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
4426 instead of X509err, which often resulted in confusing error
4427 messages since the error codes are not globally unique
4428 (e.g. an alleged error in ssl3_accept when a certificate
4429 didn't match the private key).
4431 *) New function SSL_CTX_set_session_id_context that allows to set a default
4432 value (so that you don't need SSL_set_session_id_context for each
4433 connection using the SSL_CTX).
4436 *) OAEP decoding bug fix.
4439 *) Support INSTALL_PREFIX for package builders, as proposed by
4443 *) New Configure options "threads" and "no-threads". For systems
4444 where the proper compiler options are known (currently Solaris
4445 and Linux), "threads" is the default.
4448 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
4451 *) Install various scripts to $(OPENSSLDIR)/misc, not to
4452 $(INSTALLTOP)/bin -- they shouldn't clutter directories
4453 such as /usr/local/bin.
4456 *) "make linux-shared" to build shared libraries.
4457 [Niels Poppe <niels@netbox.org>]
4459 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
4462 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
4463 extension adding in x509 utility.
4466 *) Remove NOPROTO sections and error code comments.
4469 *) Partial rewrite of the DEF file generator to now parse the ANSI
4473 *) New Configure options --prefix=DIR and --openssldir=DIR.
4476 *) Complete rewrite of the error code script(s). It is all now handled
4477 by one script at the top level which handles error code gathering,
4478 header rewriting and C source file generation. It should be much better
4479 than the old method: it now uses a modified version of Ulf's parser to
4480 read the ANSI prototypes in all header files (thus the old K&R definitions
4481 aren't needed for error creation any more) and do a better job of
4482 translating function codes into names. The old 'ASN1 error code imbedded
4483 in a comment' is no longer necessary and it doesn't use .err files which
4484 have now been deleted. Also the error code call doesn't have to appear all
4485 on one line (which resulted in some large lines...).
4488 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
4491 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
4492 0 (which usually indicates a closed connection), but continue reading.
4495 *) Fix some race conditions.
4498 *) Add support for CRL distribution points extension. Add Certificate
4499 Policies and CRL distribution points documentation.
4502 *) Move the autogenerated header file parts to crypto/opensslconf.h.
4505 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
4506 8 of keying material. Merlin has also confirmed interop with this fix
4507 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
4508 [Merlin Hughes <merlin@baltimore.ie>]
4510 *) Fix lots of warnings.
4511 [Richard Levitte <levitte@stacken.kth.se>]
4513 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
4514 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
4515 [Richard Levitte <levitte@stacken.kth.se>]
4517 *) Fix problems with sizeof(long) == 8.
4518 [Andy Polyakov <appro@fy.chalmers.se>]
4520 *) Change functions to ANSI C.
4523 *) Fix typos in error codes.
4524 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
4526 *) Remove defunct assembler files from Configure.
4529 *) SPARC v8 assembler BIGNUM implementation.
4530 [Andy Polyakov <appro@fy.chalmers.se>]
4532 *) Support for Certificate Policies extension: both print and set.
4533 Various additions to support the r2i method this uses.
4536 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
4537 return a const string when you are expecting an allocated buffer.
4540 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
4541 types DirectoryString and DisplayText.
4544 *) Add code to allow r2i extensions to access the configuration database,
4545 add an LHASH database driver and add several ctx helper functions.
4548 *) Fix an evil bug in bn_expand2() which caused various BN functions to
4549 fail when they extended the size of a BIGNUM.
4552 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
4553 support typesafe stack.
4556 *) Fix typo in SSL_[gs]et_options().
4557 [Nils Frostberg <nils@medcom.se>]
4559 *) Delete various functions and files that belonged to the (now obsolete)
4560 old X509V3 handling code.
4563 *) New Configure option "rsaref".
4566 *) Don't auto-generate pem.h.
4569 *) Introduce type-safe ASN.1 SETs.
4572 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
4573 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
4575 *) Introduce type-safe STACKs. This will almost certainly break lots of code
4576 that links with OpenSSL (well at least cause lots of warnings), but fear
4577 not: the conversion is trivial, and it eliminates loads of evil casts. A
4578 few STACKed things have been converted already. Feel free to convert more.
4579 In the fullness of time, I'll do away with the STACK type altogether.
4582 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
4583 specified in <certfile> by updating the entry in the index.txt file.
4584 This way one no longer has to edit the index.txt file manually for
4585 revoking a certificate. The -revoke option does the gory details now.
4586 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
4588 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
4589 `-text' option at all and this way the `-noout -text' combination was
4590 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
4591 [Ralf S. Engelschall]
4593 *) Make sure a corresponding plain text error message exists for the
4594 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
4595 verify callback function determined that a certificate was revoked.
4596 [Ralf S. Engelschall]
4598 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
4599 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
4600 all available cipers including rc5, which was forgotten until now.
4601 In order to let the testing shell script know which algorithms
4602 are available, a new (up to now undocumented) command
4603 "openssl list-cipher-commands" is used.
4606 *) Bugfix: s_client occasionally would sleep in select() when
4607 it should have checked SSL_pending() first.
4610 *) New functions DSA_do_sign and DSA_do_verify to provide access to
4611 the raw DSA values prior to ASN.1 encoding.
4614 *) Tweaks to Configure
4615 [Niels Poppe <niels@netbox.org>]
4617 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
4621 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
4624 *) New config option to avoid instructions that are illegal on the 80386.
4625 The default code is faster, but requires at least a 486.
4628 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
4629 SSL2_SERVER_VERSION (not used at all) macros, which are now the
4630 same as SSL2_VERSION anyway.
4633 *) New "-showcerts" option for s_client.
4636 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
4637 application. Various cleanups and fixes.
4640 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
4641 modify error routines to work internally. Add error codes and PBE init
4642 to library startup routines.
4645 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
4646 packing functions to asn1 and evp. Changed function names and error
4647 codes along the way.
4650 *) PKCS12 integration: and so it begins... First of several patches to
4651 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
4652 objects to objects.h
4655 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
4656 and display support for Thawte strong extranet extension.
4659 *) Add LinuxPPC support.
4660 [Jeff Dubrule <igor@pobox.org>]
4662 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
4663 bn_div_words in alpha.s.
4664 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
4666 *) Make sure the RSA OAEP test is skipped under -DRSAref because
4667 OAEP isn't supported when OpenSSL is built with RSAref.
4668 [Ulf Moeller <ulf@fitug.de>]
4670 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
4671 so they no longer are missing under -DNOPROTO.
4672 [Soren S. Jorvang <soren@t.dk>]
4675 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
4677 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
4678 doesn't work when the session is reused. Coming soon!
4681 *) Fix a security hole, that allows sessions to be reused in the wrong
4682 context thus bypassing client cert protection! All software that uses
4683 client certs and session caches in multiple contexts NEEDS PATCHING to
4684 allow session reuse! A fuller solution is in the works.
4685 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
4687 *) Some more source tree cleanups (removed obsolete files
4688 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
4689 permission on "config" script to be executable) and a fix for the INSTALL
4691 [Ulf Moeller <ulf@fitug.de>]
4693 *) Remove some legacy and erroneous uses of malloc, free instead of
4695 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
4697 *) Make rsa_oaep_test return non-zero on error.
4698 [Ulf Moeller <ulf@fitug.de>]
4700 *) Add support for native Solaris shared libraries. Configure
4701 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
4702 if someone would make that last step automatic.
4703 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
4705 *) ctx_size was not built with the right compiler during "make links". Fixed.
4708 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
4709 except NULL ciphers". This means the default cipher list will no longer
4710 enable NULL ciphers. They need to be specifically enabled e.g. with
4711 the string "DEFAULT:eNULL".
4714 *) Fix to RSA private encryption routines: if p < q then it would
4715 occasionally produce an invalid result. This will only happen with
4716 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
4719 *) Be less restrictive and allow also `perl util/perlpath.pl
4720 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
4721 because this way one can also use an interpreter named `perl5' (which is
4722 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
4723 installed as `perl').
4724 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
4726 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
4727 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
4729 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
4730 advapi32.lib to Win32 build and change the pem test comparision
4731 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
4732 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
4733 and crypto/des/ede_cbcm_enc.c.
4736 *) DES quad checksum was broken on big-endian architectures. Fixed.
4739 *) Comment out two functions in bio.h that aren't implemented. Fix up the
4740 Win32 test batch file so it (might) work again. The Win32 test batch file
4741 is horrible: I feel ill....
4744 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
4745 in e_os.h. Audit of header files to check ANSI and non ANSI
4746 sections: 10 functions were absent from non ANSI section and not exported
4747 from Windows DLLs. Fixed up libeay.num for new functions.
4750 *) Make `openssl version' output lines consistent.
4751 [Ralf S. Engelschall]
4753 *) Fix Win32 symbol export lists for BIO functions: Added
4754 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
4755 to ms/libeay{16,32}.def.
4756 [Ralf S. Engelschall]
4758 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
4759 fine under Unix and passes some trivial tests I've now added. But the
4760 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
4761 added to make sure no one expects that this stuff really works in the
4762 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
4763 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
4765 [Ralf S. Engelschall]
4767 *) Fix the generation of two part addresses in perl.
4768 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
4770 *) Add config entry for Linux on MIPS.
4771 [John Tobey <jtobey@channel1.com>]
4773 *) Make links whenever Configure is run, unless we are on Windoze.
4776 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
4777 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
4781 *) Add a useful kludge to allow package maintainers to specify compiler and
4782 other platforms details on the command line without having to patch the
4783 Configure script everytime: One now can use ``perl Configure
4784 <id>:<details>'', i.e. platform ids are allowed to have details appended
4785 to them (seperated by colons). This is treated as there would be a static
4786 pre-configured entry in Configure's %table under key <id> with value
4787 <details> and ``perl Configure <id>'' is called. So, when you want to
4788 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
4789 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
4790 now, which overrides the FreeBSD-elf entry on-the-fly.
4791 [Ralf S. Engelschall]
4793 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
4796 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
4797 on the `perl Configure ...' command line. This way one can compile
4798 OpenSSL libraries with Position Independent Code (PIC) which is needed
4799 for linking it into DSOs.
4800 [Ralf S. Engelschall]
4802 *) Remarkably, export ciphers were totally broken and no-one had noticed!
4806 *) Cleaned up the LICENSE document: The official contact for any license
4807 questions now is the OpenSSL core team under openssl-core@openssl.org.
4808 And add a paragraph about the dual-license situation to make sure people
4809 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
4810 to the OpenSSL toolkit.
4811 [Ralf S. Engelschall]
4813 *) General source tree makefile cleanups: Made `making xxx in yyy...'
4814 display consistent in the source tree and replaced `/bin/rm' by `rm'.
4815 Additonally cleaned up the `make links' target: Remove unnecessary
4816 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
4817 to speed processing and no longer clutter the display with confusing
4818 stuff. Instead only the actually done links are displayed.
4819 [Ralf S. Engelschall]
4821 *) Permit null encryption ciphersuites, used for authentication only. It used
4822 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
4823 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
4827 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
4828 signed attributes when verifying signatures (this would break them),
4829 the detached data encoding was wrong and public keys obtained using
4830 X509_get_pubkey() weren't freed.
4833 *) Add text documentation for the BUFFER functions. Also added a work around
4834 to a Win95 console bug. This was triggered by the password read stuff: the
4835 last character typed gets carried over to the next fread(). If you were
4836 generating a new cert request using 'req' for example then the last
4837 character of the passphrase would be CR which would then enter the first
4841 *) Added the new `Includes OpenSSL Cryptography Software' button as
4842 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
4843 button and can be used by applications based on OpenSSL to show the
4844 relationship to the OpenSSL project.
4845 [Ralf S. Engelschall]
4847 *) Remove confusing variables in function signatures in files
4848 ssl/ssl_lib.c and ssl/ssl.h.
4849 [Lennart Bong <lob@kulthea.stacken.kth.se>]
4851 *) Don't install bss_file.c under PREFIX/include/
4852 [Lennart Bong <lob@kulthea.stacken.kth.se>]
4854 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
4855 functions that return function pointers and has support for NT specific
4856 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
4857 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
4858 unsigned to signed types: this was killing the Win32 compile.
4861 *) Add new certificate file to stack functions,
4862 SSL_add_dir_cert_subjects_to_stack() and
4863 SSL_add_file_cert_subjects_to_stack(). These largely supplant
4864 SSL_load_client_CA_file(), and can be used to add multiple certs easily
4865 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
4866 This means that Apache-SSL and similar packages don't have to mess around
4867 to add as many CAs as they want to the preferred list.
4870 *) Experiment with doxygen documentation. Currently only partially applied to
4872 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
4873 openssl.doxy as the configuration file.
4876 *) Get rid of remaining C++-style comments which strict C compilers hate.
4877 [Ralf S. Engelschall, pointed out by Carlos Amengual]
4879 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
4880 compiled in by default: it has problems with large keys.
4883 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
4884 DH private keys and/or callback functions which directly correspond to
4885 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
4886 is needed for applications which have to configure certificates on a
4887 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
4889 For the RSA certificate situation is makes no difference, but
4890 for the DSA certificate situation this fixes the "no shared cipher"
4891 problem where the OpenSSL cipher selection procedure failed because the
4892 temporary keys were not overtaken from the context and the API provided
4893 no way to reconfigure them.
4894 The new functions now let applications reconfigure the stuff and they
4895 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
4896 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
4897 non-public-API function ssl_cert_instantiate() is used as a helper
4898 function and also to reduce code redundancy inside ssl_rsa.c.
4899 [Ralf S. Engelschall]
4901 *) Move s_server -dcert and -dkey options out of the undocumented feature
4902 area because they are useful for the DSA situation and should be
4903 recognized by the users.
4904 [Ralf S. Engelschall]
4906 *) Fix the cipher decision scheme for export ciphers: the export bits are
4907 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
4908 SSL_EXP_MASK. So, the original variable has to be used instead of the
4909 already masked variable.
4910 [Richard Levitte <levitte@stacken.kth.se>]
4912 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
4913 [Richard Levitte <levitte@stacken.kth.se>]
4915 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
4916 from `int' to `unsigned int' because it's a length and initialized by
4917 EVP_DigestFinal() which expects an `unsigned int *'.
4918 [Richard Levitte <levitte@stacken.kth.se>]
4920 *) Don't hard-code path to Perl interpreter on shebang line of Configure
4921 script. Instead use the usual Shell->Perl transition trick.
4922 [Ralf S. Engelschall]
4924 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
4925 (in addition to RSA certificates) to match the behaviour of `openssl dsa
4926 -noout -modulus' as it's already the case for `openssl rsa -noout
4927 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
4928 currently the public key is printed (a decision which was already done by
4929 `openssl dsa -modulus' in the past) which serves a similar purpose.
4930 Additionally the NO_RSA no longer completely removes the whole -modulus
4931 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
4933 [Ralf S. Engelschall]
4935 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
4936 BIO. See the source (crypto/evp/bio_ok.c) for more info.
4937 [Arne Ansper <arne@ats.cyber.ee>]
4939 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
4940 to be added. Now both 'req' and 'ca' can use new objects defined in the
4944 *) Add cool BIO that does syslog (or event log on NT).
4945 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
4947 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
4948 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
4949 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
4950 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
4953 *) Add preliminary config info for new extension code.
4956 *) Make RSA_NO_PADDING really use no padding.
4957 [Ulf Moeller <ulf@fitug.de>]
4959 *) Generate errors when private/public key check is done.
4962 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
4963 for some CRL extensions and new objects added.
4966 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
4967 key usage extension and fuller support for authority key id.
4970 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
4971 padding method for RSA, which is recommended for new applications in PKCS
4972 #1 v2.0 (RFC 2437, October 1998).
4973 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
4974 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
4975 against Bleichbacher's attack on RSA.
4976 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
4979 *) Updates to the new SSL compression code
4980 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
4982 *) Fix so that the version number in the master secret, when passed
4983 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
4984 (because the server will not accept higher), that the version number
4985 is 0x03,0x01, not 0x03,0x00
4986 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
4988 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
4989 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
4990 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
4993 *) Support for RAW extensions where an arbitrary extension can be
4994 created by including its DER encoding. See apps/openssl.cnf for
4998 *) Make sure latest Perl versions don't interpret some generated C array
4999 code as Perl array code in the crypto/err/err_genc.pl script.
5000 [Lars Weber <3weber@informatik.uni-hamburg.de>]
5002 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
5003 not many people have the assembler. Various Win32 compilation fixes and
5004 update to the INSTALL.W32 file with (hopefully) more accurate Win32
5008 *) Modify configure script 'Configure' to automatically create crypto/date.h
5009 file under Win32 and also build pem.h from pem.org. New script
5010 util/mkfiles.pl to create the MINFO file on environments that can't do a
5011 'make files': perl util/mkfiles.pl >MINFO should work.
5014 *) Major rework of DES function declarations, in the pursuit of correctness
5015 and purity. As a result, many evil casts evaporated, and some weirdness,
5016 too. You may find this causes warnings in your code. Zapping your evil
5017 casts will probably fix them. Mostly.
5020 *) Fix for a typo in asn1.h. Bug fix to object creation script
5021 obj_dat.pl. It considered a zero in an object definition to mean
5022 "end of object": none of the objects in objects.h have any zeros
5023 so it wasn't spotted.
5024 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
5026 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
5027 Masking (CBCM). In the absence of test vectors, the best I have been able
5028 to do is check that the decrypt undoes the encrypt, so far. Send me test
5029 vectors if you have them.
5032 *) Correct calculation of key length for export ciphers (too much space was
5033 allocated for null ciphers). This has not been tested!
5036 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
5037 message is now correct (it understands "crypto" and "ssl" on its
5038 command line). There is also now an "update" option. This will update
5039 the util/ssleay.num and util/libeay.num files with any new functions.
5041 perl util/mkdef.pl crypto ssl update
5042 it will update them.
5045 *) Overhauled the Perl interface (perl/*):
5046 - ported BN stuff to OpenSSL's different BN library
5047 - made the perl/ source tree CVS-aware
5048 - renamed the package from SSLeay to OpenSSL (the files still contain
5049 their history because I've copied them in the repository)
5050 - removed obsolete files (the test scripts will be replaced
5051 by better Test::Harness variants in the future)
5052 [Ralf S. Engelschall]
5054 *) First cut for a very conservative source tree cleanup:
5055 1. merge various obsolete readme texts into doc/ssleay.txt
5056 where we collect the old documents and readme texts.
5057 2. remove the first part of files where I'm already sure that we no
5058 longer need them because of three reasons: either they are just temporary
5059 files which were left by Eric or they are preserved original files where
5060 I've verified that the diff is also available in the CVS via "cvs diff
5061 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
5062 the crypto/md/ stuff).
5063 [Ralf S. Engelschall]
5065 *) More extension code. Incomplete support for subject and issuer alt
5066 name, issuer and authority key id. Change the i2v function parameters
5067 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
5068 what that's for :-) Fix to ASN1 macro which messed up
5069 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
5072 *) Preliminary support for ENUMERATED type. This is largely copied from the
5076 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
5077 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
5079 *) Make sure `make rehash' target really finds the `openssl' program.
5080 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
5082 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
5083 like to hear about it if this slows down other processors.
5086 *) Add CygWin32 platform information to Configure script.
5087 [Alan Batie <batie@aahz.jf.intel.com>]
5089 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
5090 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
5092 *) New program nseq to manipulate netscape certificate sequences
5095 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
5099 *) Fixes to BN code. Previously the default was to define BN_RECURSION
5100 but the BN code had some problems that would cause failures when
5101 doing certificate verification and some other functions.
5102 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
5104 *) Add ASN1 and PEM code to support netscape certificate sequences.
5107 *) Add ASN1 and PEM code to support netscape certificate sequences.
5110 *) Add several PKIX and private extended key usage OIDs.
5113 *) Modify the 'ca' program to handle the new extension code. Modify
5114 openssl.cnf for new extension format, add comments.
5117 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
5118 and add a sample to openssl.cnf so req -x509 now adds appropriate
5122 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
5123 error code, add initial support to X509_print() and x509 application.
5126 *) Takes a deep breath and start addding X509 V3 extension support code. Add
5127 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
5128 stuff is currently isolated and isn't even compiled yet.
5131 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
5132 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
5133 Removed the versions check from X509 routines when loading extensions:
5134 this allows certain broken certificates that don't set the version
5135 properly to be processed.
5138 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
5139 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
5140 can still be regenerated with "make depend".
5143 *) Spelling mistake in C version of CAST-128.
5144 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
5146 *) Changes to the error generation code. The perl script err-code.pl
5147 now reads in the old error codes and retains the old numbers, only
5148 adding new ones if necessary. It also only changes the .err files if new
5149 codes are added. The makefiles have been modified to only insert errors
5150 when needed (to avoid needlessly modifying header files). This is done
5151 by only inserting errors if the .err file is newer than the auto generated
5152 C file. To rebuild all the error codes from scratch (the old behaviour)
5153 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
5154 or delete all the .err files.
5157 *) CAST-128 was incorrectly implemented for short keys. The C version has
5158 been fixed, but is untested. The assembler versions are also fixed, but
5159 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
5160 to regenerate it if needed.
5161 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
5162 Hagino <itojun@kame.net>]
5164 *) File was opened incorrectly in randfile.c.
5165 [Ulf Möller <ulf@fitug.de>]
5167 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
5168 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
5169 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
5170 al: it's just almost always a UTCTime. Note this patch adds new error
5171 codes so do a "make errors" if there are problems.
5174 *) Correct Linux 1 recognition in config.
5175 [Ulf Möller <ulf@fitug.de>]
5177 *) Remove pointless MD5 hash when using DSA keys in ca.
5178 [Anonymous <nobody@replay.com>]
5180 *) Generate an error if given an empty string as a cert directory. Also
5181 generate an error if handed NULL (previously returned 0 to indicate an
5182 error, but didn't set one).
5183 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
5185 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
5188 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
5189 parameters. This was causing a warning which killed off the Win32 compile.
5192 *) Remove C++ style comments from crypto/bn/bn_local.h.
5193 [Neil Costigan <neil.costigan@celocom.com>]
5195 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
5196 based on a text string, looking up short and long names and finally
5197 "dot" format. The "dot" format stuff didn't work. Added new function
5198 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
5199 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
5200 OID is not part of the table.
5203 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
5204 X509_LOOKUP_by_alias().
5207 *) Sort openssl functions by name.
5210 *) Get the gendsa program working (hopefully) and add it to app list. Remove
5211 encryption from sample DSA keys (in case anyone is interested the password
5215 *) Make _all_ *_free functions accept a NULL pointer.
5216 [Frans Heymans <fheymans@isaserver.be>]
5218 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
5220 [Anonymous <nobody@replay.com>]
5222 *) s_server should send the CAfile as acceptable CAs, not its own cert.
5223 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
5225 *) Don't blow it for numeric -newkey arguments to apps/req.
5226 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
5228 *) Temp key "for export" tests were wrong in s3_srvr.c.
5229 [Anonymous <nobody@replay.com>]
5231 *) Add prototype for temp key callback functions
5232 SSL_CTX_set_tmp_{rsa,dh}_callback().
5235 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
5236 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
5239 *) X509_name_add_entry() freed the wrong thing after an error.
5240 [Arne Ansper <arne@ats.cyber.ee>]
5242 *) rsa_eay.c would attempt to free a NULL context.
5243 [Arne Ansper <arne@ats.cyber.ee>]
5245 *) BIO_s_socket() had a broken should_retry() on Windoze.
5246 [Arne Ansper <arne@ats.cyber.ee>]
5248 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
5249 [Arne Ansper <arne@ats.cyber.ee>]
5251 *) Make sure the already existing X509_STORE->depth variable is initialized
5252 in X509_STORE_new(), but document the fact that this variable is still
5253 unused in the certificate verification process.
5254 [Ralf S. Engelschall]
5256 *) Fix the various library and apps files to free up pkeys obtained from
5257 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
5260 *) Fix reference counting in X509_PUBKEY_get(). This makes
5261 demos/maurice/example2.c work, amongst others, probably.
5262 [Steve Henson and Ben Laurie]
5264 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
5265 `openssl' and second, the shortcut symlinks for the `openssl <command>'
5266 are no longer created. This way we have a single and consistent command
5267 line interface `openssl <command>', similar to `cvs <command>'.
5268 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
5270 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
5271 BIT STRING wrapper always have zero unused bits.
5274 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
5277 *) Make the top-level INSTALL documentation easier to understand.
5280 *) Makefiles updated to exit if an error occurs in a sub-directory
5281 make (including if user presses ^C) [Paul Sutton]
5283 *) Make Montgomery context stuff explicit in RSA data structure.
5286 *) Fix build order of pem and err to allow for generated pem.h.
5289 *) Fix renumbering bug in X509_NAME_delete_entry().
5292 *) Enhanced the err-ins.pl script so it makes the error library number
5293 global and can add a library name. This is needed for external ASN1 and
5294 other error libraries.
5297 *) Fixed sk_insert which never worked properly.
5300 *) Fix ASN1 macros so they can handle indefinite length construted
5301 EXPLICIT tags. Some non standard certificates use these: they can now
5305 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
5306 into a single doc/ssleay.txt bundle. This way the information is still
5307 preserved but no longer messes up this directory. Now it's new room for
5308 the new set of documenation files.
5309 [Ralf S. Engelschall]
5311 *) SETs were incorrectly DER encoded. This was a major pain, because they
5312 shared code with SEQUENCEs, which aren't coded the same. This means that
5313 almost everything to do with SETs or SEQUENCEs has either changed name or
5314 number of arguments.
5315 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
5317 *) Fix test data to work with the above.
5320 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
5321 was already fixed by Eric for 0.9.1 it seems.
5322 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
5324 *) Autodetect FreeBSD3.
5327 *) Fix various bugs in Configure. This affects the following platforms:
5331 unixware-2.0-pentium
5335 *) Eliminate generated files from CVS. Reorder tests to regenerate files
5336 before they are needed.
5339 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
5343 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
5345 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
5346 changed SSLeay to OpenSSL in version strings.
5347 [Ralf S. Engelschall]
5349 *) Some fixups to the top-level documents.
5352 *) Fixed the nasty bug where rsaref.h was not found under compile-time
5353 because the symlink to include/ was missing.
5354 [Ralf S. Engelschall]
5356 *) Incorporated the popular no-RSA/DSA-only patches
5357 which allow to compile a RSA-free SSLeay.
5358 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
5360 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
5361 when "ssleay" is still not found.
5362 [Ralf S. Engelschall]
5364 *) Added more platforms to Configure: Cray T3E, HPUX 11,
5365 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
5367 *) Updated the README file.
5368 [Ralf S. Engelschall]
5370 *) Added various .cvsignore files in the CVS repository subdirs
5371 to make a "cvs update" really silent.
5372 [Ralf S. Engelschall]
5374 *) Recompiled the error-definition header files and added
5375 missing symbols to the Win32 linker tables.
5376 [Ralf S. Engelschall]
5378 *) Cleaned up the top-level documents;
5379 o new files: CHANGES and LICENSE
5380 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
5381 o merged COPYRIGHT into LICENSE
5382 o removed obsolete TODO file
5383 o renamed MICROSOFT to INSTALL.W32
5384 [Ralf S. Engelschall]
5386 *) Removed dummy files from the 0.9.1b source tree:
5387 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
5388 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
5389 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
5390 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
5391 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
5392 [Ralf S. Engelschall]
5394 *) Added various platform portability fixes.
5397 *) The Genesis of the OpenSSL rpject:
5398 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
5399 Young and Tim J. Hudson created while they were working for C2Net until
5401 [The OpenSSL Project]
5404 Changes between 0.9.0b and 0.9.1b [not released]
5406 *) Updated a few CA certificates under certs/
5409 *) Changed some BIGNUM api stuff.
5412 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
5413 DGUX x86, Linux Alpha, etc.
5416 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
5417 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
5421 *) Add -strparse option to asn1pars program which parses nested
5423 [Dr Stephen Henson <shenson@bigfoot.com>]
5425 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
5428 *) DSA fix for "ca" program.
5431 *) Added "-genkey" option to "dsaparam" program.
5434 *) Added RIPE MD160 (rmd160) message digest.
5437 *) Added -a (all) option to "ssleay version" command.
5440 *) Added PLATFORM define which is the id given to Configure.
5443 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
5446 *) Extended the ASN.1 parser routines.
5449 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
5452 *) Added a BN_CTX to the BN library.
5455 *) Fixed the weak key values in DES library
5458 *) Changed API in EVP library for cipher aliases.
5461 *) Added support for RC2/64bit cipher.
5464 *) Converted the lhash library to the crypto/mem.c functions.
5467 *) Added more recognized ASN.1 object ids.
5470 *) Added more RSA padding checks for SSL/TLS.
5473 *) Added BIO proxy/filter functionality.
5476 *) Added extra_certs to SSL_CTX which can be used
5477 send extra CA certificates to the client in the CA cert chain sending
5478 process. It can be configured with SSL_CTX_add_extra_chain_cert().
5481 *) Now Fortezza is denied in the authentication phase because
5482 this is key exchange mechanism is not supported by SSLeay at all.
5485 *) Additional PKCS1 checks.
5488 *) Support the string "TLSv1" for all TLS v1 ciphers.
5491 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
5492 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
5495 *) Fixed a few memory leaks.
5498 *) Fixed various code and comment typos.
5501 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
5502 bytes sent in the client random.
5503 [Edward Bishop <ebishop@spyglass.com>]