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 +) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
16 of the e-mail address in the DN (i.e., it will go into a certificate
17 extension only). The new configuration file option 'email_in_dn = no'
19 [Massimiliano Pala madwolf@openca.org]
21 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
22 faced with a pathologically small ClientHello fragment that does
23 not contain client_version: Instead of aborting with an error,
24 simply choose the highest available protocol version (i.e.,
25 TLS 1.0 unless it is disabled). In practice, ClientHello
26 messages are never sent like this, but this change gives us
27 strictly correct behaviour at least for TLS.
30 +) Change all functions with names starting with des_ to be starting
31 with DES_ instead. This because there are increasing clashes with
32 libdes and other des libraries that are currently used by other
33 projects. The old libdes interface is provided, as well as crypt(),
34 if openssl/des_old.h is included. Note that crypt() is no longer
35 declared in openssl/des.h.
37 NOTE: This is a major break of an old API into a new one. Software
38 authors are encouraged to switch to the DES_ style functions. Some
39 time in the future, des_old.h and the libdes compatibility functions
40 will be completely removed.
43 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
44 never resets s->method to s->ctx->method when called from within
45 one of the SSL handshake functions.
46 [Bodo Moeller; problem pointed out by Niko Baric]
48 +) Test for certificates which contain unsupported critical extensions.
49 If such a certificate is found during a verify operation it is
50 rejected by default: this behaviour can be overridden by either
51 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
52 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
53 X509_supported_extension() has also been added which returns 1 if a
54 particular extension is supported.
57 +) New functions/macros
59 SSL_CTX_set_msg_callback(ctx, cb)
60 SSL_CTX_set_msg_callback_arg(ctx, arg)
61 SSL_set_msg_callback(ssl, cb)
62 SSL_set_msg_callback_arg(ssl, arg)
64 to request calling a callback function
66 void cb(int write_p, int version, int content_type,
67 const void *buf, size_t len, SSL *ssl, void *arg)
69 whenever a protocol message has been completely received
70 (write_p == 0) or sent (write_p == 1). Here 'version' is the
71 protocol version according to which the SSL library interprets
72 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
73 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
74 the content type as defined in the SSL 3.0/TLS 1.0 protocol
75 specification (change_cipher_spec(20), alert(21), handshake(22)).
76 'buf' and 'len' point to the actual message, 'ssl' to the
77 SSL object, and 'arg' is the application-defined value set by
78 SSL[_CTX]_set_msg_callback_arg().
80 'openssl s_client' and 'openssl s_server' have new '-msg' options
81 to enable a callback that displays all protocol messages.
83 TODO: SSL 2.0, doc/ssl/, doc/apps/
86 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
87 (sent using the client's version number) if client_version is
88 smaller than the protocol version in use. Also change
89 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
90 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
91 the client will at least see that alert.
94 +) Modify the behaviour of EVP cipher functions in similar way to digests
95 to retain compatibility with existing code.
98 +) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
99 compatibility with existing code. In particular the 'ctx' parameter does
100 not have to be to be initialized before the call to EVP_DigestInit() and
101 it is tidied up after a call to EVP_DigestFinal(). New function
102 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
103 EVP_MD_CTX_copy() changed to not require the destination to be
104 initialized valid and new function EVP_MD_CTX_copy_ex() added which
105 requires the destination to be valid.
107 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
108 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
111 +) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
112 so that complete 'Handshake' protocol structures are kept in memory
113 instead of overwriting 'msg_type' and 'length' with 'body' data.
116 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
120 +) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
121 [Massimo Santin via Richard Levitte]
123 +) Major restructuring to the underlying ENGINE code. This includes
124 reduction of linker bloat, separation of pure "ENGINE" manipulation
125 (initialisation, etc) from functionality dealing with implementations
126 of specific crypto iterfaces. This change also introduces integrated
127 support for symmetric ciphers and digest implementations - so ENGINEs
128 can now accelerate these by providing EVP_CIPHER and EVP_MD
129 implementations of their own. This is detailed in crypto/engine/README
130 as it couldn't be adequately described here. However, there are a few
131 API changes worth noting - some RSA, DSA, DH, and RAND functions that
132 were changed in the original introduction of ENGINE code have now
133 reverted back - the hooking from this code to ENGINE is now a good
134 deal more passive and at run-time, operations deal directly with
135 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
136 dereferencing through an ENGINE pointer any more. Also, the ENGINE
137 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
138 they were not being used by the framework as there is no concept of a
139 BIGNUM_METHOD and they could not be generalised to the new
140 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
141 ENGINE_cpy() has been removed as it cannot be consistently defined in
145 +) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
148 +) Change mkdef.pl to sort symbols that get the same entry number,
149 and make sure the automatically generated functions ERR_load_*
150 become part of libeay.num as well.
153 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
154 client receives HelloRequest while in a handshake.
155 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
157 +) New function SSL_renegotiate_pending(). This returns true once
158 renegotiation has been requested (either SSL_renegotiate() call
159 or HelloRequest/ClientHello receveived from the peer) and becomes
160 false once a handshake has been completed.
161 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
162 sends a HelloRequest, but does not ensure that a handshake takes
163 place. SSL_renegotiate_pending() is useful for checking if the
164 client has followed the request.)
167 +) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
168 By default, clients may request session resumption even during
169 renegotiation (if session ID contexts permit); with this option,
170 session resumption is possible only in the first handshake.
173 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
174 should end in 'break', not 'goto end' which circuments various
175 cleanups done in state SSL_ST_OK. But session related stuff
176 must be disabled for SSL_ST_OK in the case that we just sent a
179 Also avoid some overhead by not calling ssl_init_wbio_buffer()
180 before just sending a HelloRequest.
181 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
183 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
184 reveal whether illegal block cipher padding was found or a MAC
185 verification error occured. (Neither SSLerr() codes nor alerts
186 are directly visible to potential attackers, but the information
187 may leak via logfiles.)
189 Similar changes are not required for the SSL 2.0 implementation
190 because the number of padding bytes is sent in clear for SSL 2.0,
191 and the extra bytes are just ignored. However ssl/s2_pkt.c
192 failed to verify that the purported number of padding bytes is in
196 +) Add some demos for certificate and certificate request creation.
199 +) Make maximum certificate chain size accepted from the peer application
200 settable (SSL*_get/set_max_cert_list()), as proposed by
201 "Douglas E. Engert" <deengert@anl.gov>.
204 +) Add support for shared libraries for Unixware-7 and support including
205 shared libraries for OpenUNIX-8 (Boyd Lynn Gerber <gerberb@zenez.com>).
208 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
209 'wristwatch attack' using huge encoding parameters (cf.
210 James H. Manger's CRYPTO 2001 paper). Note that the
211 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
212 encoding parameters and hence was not vulnerable.
215 +) Add a "destroy" handler to ENGINEs that allows structural cleanup to
216 be done prior to destruction. Use this to unload error strings from
217 ENGINEs that load their own error strings. NB: This adds two new API
218 functions to "get" and "set" this destroy handler in an ENGINE.
221 +) Alter all existing ENGINE implementations (except "openssl" and
222 "openbsd") to dynamically instantiate their own error strings. This
223 makes them more flexible to be built both as statically-linked ENGINEs
224 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
225 Also, add stub code to each that makes building them as self-contained
226 shared-libraries easier (see README.ENGINE).
229 +) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
230 implementations into applications that are completely implemented in
231 self-contained shared-libraries. The "dynamic" ENGINE exposes control
232 commands that can be used to configure what shared-library to load and
233 to control aspects of the way it is handled. Also, made an update to
234 the README.ENGINE file that brings its information up-to-date and
235 provides some information and instructions on the "dynamic" ENGINE
236 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
240 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
242 *) Make it possible to unload ranges of ERR strings with a new
243 "ERR_unload_strings" function.
246 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
247 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
248 followed by modular reduction.
249 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
251 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
252 equivalent based on BN_pseudo_rand() instead of BN_rand().
255 +) Add a copy() function to EVP_MD.
258 +) Make EVP_MD routines take a context pointer instead of just the
259 md_data void pointer.
262 +) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
263 that the digest can only process a single chunk of data
264 (typically because it is provided by a piece of
265 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
266 is only going to provide a single chunk of data, and hence the
267 framework needn't accumulate the data for oneshot drivers.
270 +) As with "ERR", make it possible to replace the underlying "ex_data"
271 functions. This change also alters the storage and management of global
272 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
273 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
274 index counters. The API functions that use this state have been changed
275 to take a "class_index" rather than pointers to the class's local STACK
276 and counter, and there is now an API function to dynamically create new
277 classes. This centralisation allows us to (a) plug a lot of the
278 thread-safety problems that existed, and (b) makes it possible to clean
279 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
280 such data would previously have always leaked in application code and
281 workarounds were in place to make the memory debugging turn a blind eye
282 to it. Application code that doesn't use this new function will still
283 leak as before, but their memory debugging output will announce it now
284 rather than letting it slide.
286 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
287 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
288 has a return value to indicate success or failure.
291 +) Make it possible to replace the underlying "ERR" functions such that the
292 global state (2 LHASH tables and 2 locks) is only used by the "default"
293 implementation. This change also adds two functions to "get" and "set"
294 the implementation prior to it being automatically set the first time
295 any other ERR function takes place. Ie. an application can call "get",
296 pass the return value to a module it has just loaded, and that module
297 can call its own "set" function using that value. This means the
298 module's "ERR" operations will use (and modify) the error state in the
299 application and not in its own statically linked copy of OpenSSL code.
302 +) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
303 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
304 the operation, and provides a more encapsulated way for external code
305 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
306 to use these functions rather than manually incrementing the counts.
308 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
311 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
312 This function was broken, as the check for a new client hello message
313 to handle SGC did not allow these large messages.
314 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
317 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
320 +) Add EVP test program.
323 +) Add symmetric cipher support to ENGINE. Expect the API to change!
326 +) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
327 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
328 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
329 These allow a CRL to be built without having to access X509_CRL fields
330 directly. Modify 'ca' application to use new functions.
333 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
334 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
337 *) Rework the configuration and shared library support for Tru64 Unix.
338 The configuration part makes use of modern compiler features and
339 still retains old compiler behavior for those that run older versions
340 of the OS. The shared library support part includes a variant that
341 uses the RPATH feature, and is available through the special
342 configuration target "alpha-cc-rpath", which will never be selected
344 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
346 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
347 with the same message size as in ssl3_get_certificate_request().
348 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
349 messages might inadvertently be reject as too long.
350 [Petr Lampa <lampa@fee.vutbr.cz>]
352 +) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
353 bug workarounds. Rollback attack detection is a security feature.
354 The problem will only arise on OpenSSL servers when TLSv1 is not
355 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
356 Software authors not wanting to support TLSv1 will have special reasons
357 for their choice and can explicitly enable this option.
358 [Bodo Moeller, Lutz Jaenicke]
360 +) Rationalise EVP so it can be extended: don't include a union of
361 cipher/digest structures, add init/cleanup functions. This also reduces
362 the number of header dependencies.
365 +) Make DES key schedule conform to the usual scheme, as well as
366 correcting its structure. This means that calls to DES functions
367 now have to pass a pointer to a des_key_schedule instead of a
368 plain des_key_schedule (which was actually always a pointer
372 +) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
375 *) Modified SSL library such that the verify_callback that has been set
376 specificly for an SSL object with SSL_set_verify() is actually being
377 used. Before the change, a verify_callback set with this function was
378 ignored and the verify_callback() set in the SSL_CTX at the time of
379 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
380 to allow the necessary settings.
383 +) Initial reduction of linker bloat: the use of some functions, such as
384 PEM causes large amounts of unused functions to be linked in due to
385 poor organisation. For example pem_all.c contains every PEM function
386 which has a knock on effect of linking in large amounts of (unused)
387 ASN1 code. Grouping together similar functions and splitting unrelated
388 functions prevents this.
391 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
392 explicitly to NULL, as at least on Solaris 8 this seems not always to be
393 done automatically (in contradiction to the requirements of the C
394 standard). This made problems when used from OpenSSH.
397 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
398 dh->length and always used
400 BN_rand_range(priv_key, dh->p).
402 BN_rand_range() is not necessary for Diffie-Hellman, and this
403 specific range makes Diffie-Hellman unnecessarily inefficient if
404 dh->length (recommended exponent length) is much smaller than the
405 length of dh->p. We could use BN_rand_range() if the order of
406 the subgroup was stored in the DH structure, but we only have
411 BN_rand(priv_key, l, ...)
413 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
419 RSA_eay_public_encrypt
420 RSA_eay_private_decrypt
421 RSA_eay_private_encrypt (signing)
422 RSA_eay_public_decrypt (signature verification)
424 (default implementations for RSA_public_encrypt,
425 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
426 always reject numbers >= n.
429 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
430 to synchronize access to 'locking_thread'. This is necessary on
431 systems where access to 'locking_thread' (an 'unsigned long'
432 variable) is not atomic.
435 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
436 *before* setting the 'crypto_lock_rand' flag. The previous code had
437 a race condition if 0 is a valid thread ID.
438 [Travis Vitek <vitek@roguewave.com>]
440 +) Cleanup of EVP macros.
443 +) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
447 +) Add initial OCSP responder support to ocsp application. The
448 revocation information is handled using the text based index
449 use by the ca application. The responder can either handle
450 requests generated internally, supplied in files (for example
451 via a CGI script) or using an internal minimal server.
454 +) Add configuration choices to get zlib compression for TLS.
457 +) Changes to Kerberos SSL for RFC 2712 compliance:
458 1. Implemented real KerberosWrapper, instead of just using
459 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
460 2. Implemented optional authenticator field of KerberosWrapper.
462 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
463 and authenticator structs; see crypto/krb5/.
465 Generalized Kerberos calls to support multiple Kerberos libraries.
466 [Vern Staats <staatsvr@asc.hpc.mil>,
467 Jeffrey Altman <jaltman@columbia.edu>
470 +) Cause 'openssl speed' to use fully hard-coded DSA keys as it
471 already does with RSA. testdsa.h now has 'priv_key/pub_key'
472 values for each of the key sizes rather than having just
473 parameters (and 'speed' generating keys each time).
476 -) OpenSSL 0.9.6b released [9 July 2001]
478 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
479 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
480 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
481 PRNG state recovery was possible based on the output of
482 one PRNG request appropriately sized to gain knowledge on
483 'md' followed by enough consecutive 1-byte PRNG requests
484 to traverse all of 'state'.
486 1. When updating 'md_local' (the current thread's copy of 'md')
487 during PRNG output generation, hash all of the previous
488 'md_local' value, not just the half used for PRNG output.
490 2. Make the number of bytes from 'state' included into the hash
491 independent from the number of PRNG bytes requested.
493 The first measure alone would be sufficient to avoid
494 Markku-Juhani's attack. (Actually it had never occurred
495 to me that the half of 'md_local' used for chaining was the
496 half from which PRNG output bytes were taken -- I had always
497 assumed that the secret half would be used.) The second
498 measure makes sure that additional data from 'state' is never
499 mixed into 'md_local' in small portions; this heuristically
500 further strengthens the PRNG.
503 +) Speed up EVP routines.
506 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
507 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
508 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
509 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
511 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
512 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
513 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
516 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
518 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
521 *) Fix crypto/bn/asm/mips3.s.
524 *) When only the key is given to "enc", the IV is undefined. Print out
525 an error message in this case.
528 +) Added the OS2-EMX target.
529 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
531 +) Rewrite apps to use NCONF routines instead of the old CONF. New functions
532 to support NCONF routines in extension code. New function CONF_set_nconf()
533 to allow functions which take an NCONF to also handle the old LHASH
534 structure: this means that the old CONF compatible routines can be
535 retained (in particular wrt extensions) without having to duplicate the
536 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
539 *) Handle special case when X509_NAME is empty in X509 printing routines.
542 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
543 positive and less than q.
546 +) Enhance the general user interface with mechanisms for inner control
547 and with possibilities to have yes/no kind of prompts.
550 +) Change all calls to low level digest routines in the library and
551 applications to use EVP. Add missing calls to HMAC_cleanup() and
552 don't assume HMAC_CTX can be copied using memcpy().
553 [Verdon Walker <VWalker@novell.com>, Steve Henson]
555 +) Add the possibility to control engines through control names but with
556 arbitrary arguments instead of just a string.
557 Change the key loaders to take a UI_METHOD instead of a callback
558 function pointer. NOTE: this breaks binary compatibility with earlier
559 versions of OpenSSL [engine].
560 Adapt the nCipher code for these new conditions and add a card insertion
564 +) Enhance the general user interface with mechanisms to better support
565 dialog box interfaces, application-defined prompts, the possibility
566 to use defaults (for example default passwords from somewhere else)
567 and interrupts/cancellations.
570 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
571 used: it isn't thread safe and the add_lock_callback should handle
573 [Paul Rose <Paul.Rose@bridge.com>]
575 *) Verify that incoming data obeys the block size in
576 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
579 +) Tidy up PKCS#12 attribute handling. Add support for the CSP name
580 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
584 [Ulf Möller, Bodo Möller]
586 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
587 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
588 when fixing the server behaviour for backwards-compatible 'client
589 hello' messages. (Note that the attack is impractical against
590 SSL 3.0 and TLS 1.0 anyway because length and version checking
591 means that the probability of guessing a valid ciphertext is
592 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
595 Before 0.9.5, the countermeasure (hide the error by generating a
596 random 'decryption result') did not work properly because
597 ERR_clear_error() was missing, meaning that SSL_get_error() would
598 detect the supposedly ignored error.
600 Both problems are now fixed.
603 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
604 (previously it was 1024).
607 +) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
608 tidy up some unnecessarily weird code in 'sk_new()').
609 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
611 +) Change the key loading routines for ENGINEs to use the same kind
612 callback (pem_password_cb) as all other routines that need this
616 *) Fix for compatibility mode trust settings: ignore trust settings
617 unless some valid trust or reject settings are present.
620 *) Fix for blowfish EVP: its a variable length cipher.
623 +) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
624 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
625 than this minimum value is recommended.
628 +) New random seeder for OpenVMS, using the system process statistics
629 that are easily reachable.
632 +) Windows apparently can't transparently handle global
633 variables defined in DLLs. Initialisations such as:
635 const ASN1_ITEM *it = &ASN1_INTEGER_it;
637 wont compile. This is used by the any applications that need to
638 declare their own ASN1 modules. This was fixed by adding the option
639 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
640 needed for static libraries under Win32.
643 +) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
644 setting of purpose and trust fields. New X509_STORE trust and
645 purpose functions and tidy up setting in other SSL functions.
648 +) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
649 structure. These are inherited by X509_STORE_CTX when it is
650 initialised. This allows various defaults to be set in the
651 X509_STORE structure (such as flags for CRL checking and custom
652 purpose or trust settings) for functions which only use X509_STORE_CTX
653 internally such as S/MIME.
655 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
656 trust settings if they are not set in X509_STORE. This allows X509_STORE
657 purposes and trust (in S/MIME for example) to override any set by default.
659 Add command line options for CRL checking to smime, s_client and s_server
663 +) Initial CRL based revocation checking. If the CRL checking flag(s)
664 are set then the CRL is looked up in the X509_STORE structure and
665 its validity and signature checked, then if the certificate is found
666 in the CRL the verify fails with a revoked error.
668 Various new CRL related callbacks added to X509_STORE_CTX structure.
670 Command line options added to 'verify' application to support this.
672 This needs some additional work, such as being able to handle multiple
673 CRLs with different times, extension based lookup (rather than just
674 by subject name) and ultimately more complete V2 CRL extension
678 +) Add a general user interface API (crypto/ui/). This is designed
679 to replace things like des_read_password and friends (backward
680 compatibility functions using this new API are provided).
681 The purpose is to remove prompting functions from the DES code
682 section as well as provide for prompting through dialog boxes in
683 a window system and the like.
686 *) Fix various bugs related to DSA S/MIME verification. Handle missing
687 parameters in DSA public key structures and return an error in the
688 DSA routines if parameters are absent.
691 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
692 in the current directory if neither $RANDFILE nor $HOME was set.
693 RAND_file_name() in 0.9.6a returned NULL in this case. This has
694 caused some confusion to Windows users who haven't defined $HOME.
695 Thus RAND_file_name() is changed again: e_os.h can define a
696 DEFAULT_HOME, which will be used if $HOME is not set.
697 For Windows, we use "C:"; on other platforms, we still require
698 environment variables.
700 +) Add "ex_data" support to ENGINE so implementations can add state at a
701 per-structure level rather than having to store it globally.
704 +) Make it possible for ENGINE structures to be copied when retrieved by
705 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
706 This causes the "original" ENGINE structure to act like a template,
707 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
708 operational state can be localised to each ENGINE structure, despite the
709 fact they all share the same "methods". New ENGINE structures returned in
710 this case have no functional references and the return value is the single
711 structural reference. This matches the single structural reference returned
712 by ENGINE_by_id() normally, when it is incremented on the pre-existing
716 +) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
717 needs to match any other type at all we need to manually clear the
721 +) Changes to the "openssl engine" utility to include;
722 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
723 about an ENGINE's available control commands.
724 - executing control commands from command line arguments using the
725 '-pre' and '-post' switches. '-post' is only used if '-t' is
726 specified and the ENGINE is successfully initialised. The syntax for
727 the individual commands are colon-separated, for example;
728 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
731 +) New dynamic control command support for ENGINEs. ENGINEs can now
732 declare their own commands (numbers), names (strings), descriptions,
733 and input types for run-time discovery by calling applications. A
734 subset of these commands are implicitly classed as "executable"
735 depending on their input type, and only these can be invoked through
736 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
737 can be based on user input, config files, etc). The distinction is
738 that "executable" commands cannot return anything other than a boolean
739 result and can only support numeric or string input, whereas some
740 discoverable commands may only be for direct use through
741 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
742 pointers, or other custom uses. The "executable" commands are to
743 support parameterisations of ENGINE behaviour that can be
744 unambiguously defined by ENGINEs and used consistently across any
745 OpenSSL-based application. Commands have been added to all the
746 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
747 control over shared-library paths without source code alterations.
750 +) Changed all ENGINE implementations to dynamically allocate their
751 ENGINEs rather than declaring them statically. Apart from this being
752 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
753 this also allows the implementations to compile without using the
754 internal engine_int.h header.
757 +) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
758 'const' value. Any code that should be able to modify a RAND_METHOD
759 should already have non-const pointers to it (ie. they should only
760 modify their own ones).
763 +) Made a variety of little tweaks to the ENGINE code.
764 - "atalla" and "ubsec" string definitions were moved from header files
765 to C code. "nuron" string definitions were placed in variables
766 rather than hard-coded - allowing parameterisation of these values
767 later on via ctrl() commands.
768 - Removed unused "#if 0"'d code.
769 - Fixed engine list iteration code so it uses ENGINE_free() to release
770 structural references.
771 - Constified the RAND_METHOD element of ENGINE structures.
772 - Constified various get/set functions as appropriate and added
773 missing functions (including a catch-all ENGINE_cpy that duplicates
774 all ENGINE values onto a new ENGINE except reference counts/state).
775 - Removed NULL parameter checks in get/set functions. Setting a method
776 or function to NULL is a way of cancelling out a previously set
777 value. Passing a NULL ENGINE parameter is just plain stupid anyway
778 and doesn't justify the extra error symbols and code.
779 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
780 flags from engine_int.h to engine.h.
781 - Changed prototypes for ENGINE handler functions (init(), finish(),
782 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
785 *) Move 'if (!initialized) RAND_poll()' into regions protected by
786 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
787 having multiple threads call RAND_poll() concurrently.
790 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
791 combination of a flag and a thread ID variable.
792 Otherwise while one thread is in ssleay_rand_bytes (which sets the
793 flag), *other* threads can enter ssleay_add_bytes without obeying
794 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
795 that they do not hold after the first thread unsets add_do_not_lock).
798 +) Implement binary inversion algorithm for BN_mod_inverse in addition
799 to the algorithm using long division. The binary algorithm can be
800 used only if the modulus is odd. On 32-bit systems, it is faster
801 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
802 roughly 5-15% for 256-bit moduli), so we use it only for moduli
803 up to 450 bits. In 64-bit environments, the binary algorithm
804 appears to be advantageous for much longer moduli; here we use it
805 for moduli up to 2048 bits.
808 *) Change bctest again: '-x' expressions are not available in all
812 -) OpenSSL 0.9.6a released [5 Apr 2001]
814 *) Fix a couple of memory leaks in PKCS7_dataDecode()
815 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
817 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
818 the default extension for executables, if any. Also, make the perl
819 scripts that use symlink() to test if it really exists and use "cp"
820 if it doesn't. All this made OpenSSL compilable and installable in
824 +) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
825 could not support the combine flag in choice fields.
828 -) Fix for asn1_GetSequence() for indefinite length constructed data.
829 If SEQUENCE is length is indefinite just set c->slen to the total
830 amount of data available.
831 [Steve Henson, reported by shige@FreeBSD.org]
832 [This change does not apply to 0.9.7.]
834 *) Change bctest to avoid here-documents inside command substitution
835 (workaround for FreeBSD /bin/sh bug).
836 For compatibility with Ultrix, avoid shell functions (introduced
837 in the bctest version that searches along $PATH).
840 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
841 with des_encrypt() defined on some operating systems, like Solaris
845 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
846 On the Importance of Eliminating Errors in Cryptographic
847 Computations, J. Cryptology 14 (2001) 2, 101-119,
848 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
851 *) MIPS assembler BIGNUM division bug fix.
854 *) Disabled incorrect Alpha assembler code.
857 -) Fix PKCS#7 decode routines so they correctly update the length
858 after reading an EOC for the EXPLICIT tag.
860 [This change does not apply to 0.9.7.]
862 *) Fix bug in PKCS#12 key generation routines. This was triggered
863 if a 3DES key was generated with a 0 initial byte. Include
864 PKCS12_BROKEN_KEYGEN compilation option to retain the old
865 (but broken) behaviour.
868 *) Enhance bctest to search for a working bc along $PATH and print
870 [Tim Rice <tim@multitalents.net> via Richard Levitte]
872 +) Add a 'copy_extensions' option to the 'ca' utility. This copies
873 extensions from a certificate request to the certificate.
876 +) Allow multiple 'certopt' and 'nameopt' options to be separated
877 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
878 file: this allows the display of the certificate about to be
879 signed to be customised, to allow certain fields to be included
880 or excluded and extension details. The old system didn't display
881 multicharacter strings properly, omitted fields not in the policy
882 and couldn't display additional details such as extensions.
885 *) Fix memory leaks in err.c: free err_data string if necessary;
886 don't write to the wrong index in ERR_set_error_data.
889 +) Function EC_POINTs_mul for simultaneous scalar multiplication
890 of an arbitrary number of elliptic curve points, optionally
891 including the generator defined for the EC_GROUP.
892 EC_POINT_mul is a simple wrapper function for the typical case
893 that the point list has just one item (besides the optional
897 +) First EC_METHODs for curves over GF(p):
899 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
900 operations and provides various method functions that can also
901 operate with faster implementations of modular arithmetic.
903 EC_GFp_mont_method() reuses most functions that are part of
904 EC_GFp_simple_method, but uses Montgomery arithmetic.
906 [Bodo Moeller; point addition and point doubling
907 implementation directly derived from source code provided by
908 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
910 +) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
913 Curves are EC_GROUP objects (with an optional group generator)
914 based on EC_METHODs that are built into the library.
916 Points are EC_POINT objects based on EC_GROUP objects.
918 Most of the framework would be able to handle curves over arbitrary
919 finite fields, but as there are no obvious types for fields other
920 than GF(p), some functions are limited to that for now.
923 +) Add the -HTTP option to s_server. It is similar to -WWW, but requires
924 that the file contains a complete HTTP response.
927 +) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
928 change the def and num file printf format specifier from "%-40sXXX"
929 to "%-39s XXX". The latter will always guarantee a space after the
930 field while the former will cause them to run together if the field
931 is 40 of more characters long.
934 +) Constify the cipher and digest 'method' functions and structures
935 and modify related functions to take constant EVP_MD and EVP_CIPHER
939 *) Implement ssl23_peek (analogous to ssl23_read), which previously
943 *) Replace rdtsc with _emit statements for VC++ version 5.
944 [Jeremy Cooper <jeremy@baymoo.org>]
946 +) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
947 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
950 +) Modify EVP_Digest*() routines so they now return values. Although the
951 internal software routines can never fail additional hardware versions
955 +) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
957 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
958 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
961 ERR_R_NESTED_ASN1_ERROR
963 ERR_R_MISSING_ASN1_EOS
964 were 4 .. 9, conflicting with
965 ERR_LIB_RSA (= ERR_R_RSA_LIB)
967 ERR_LIB_PEM (= ERR_R_PEM_LIB).
968 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
970 Add new error code 'ERR_R_INTERNAL_ERROR'.
973 +) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
977 +) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
978 sets the subject name for a new request or supersedes the
979 subject name in a given request. Formats that can be parsed are
980 'CN=Some Name, OU=myOU, C=IT'
982 'CN=Some Name/OU=myOU/C=IT'.
984 Add options '-batch' and '-verbose' to 'openssl req'.
985 [Massimiliano Pala <madwolf@hackmasters.net>]
987 *) Make it possible to reuse SSLv2 sessions.
990 +) Introduce the possibility to access global variables through
991 functions on platform were that's the best way to handle exporting
992 global variables in shared libraries. To enable this functionality,
993 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
994 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
995 is normally done by Configure or something similar).
997 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
998 in the source file (foo.c) like this:
1000 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
1001 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
1003 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
1004 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
1006 OPENSSL_DECLARE_GLOBAL(int,foo);
1007 #define foo OPENSSL_GLOBAL_REF(foo)
1008 OPENSSL_DECLARE_GLOBAL(double,bar);
1009 #define bar OPENSSL_GLOBAL_REF(bar)
1011 The #defines are very important, and therefore so is including the
1012 header file everywhere where the defined globals are used.
1014 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
1015 of ASN.1 items, but that structure is a bit different.
1017 The largest change is in util/mkdef.pl which has been enhanced with
1018 better and easier to understand logic to choose which symbols should
1019 go into the Windows .def files as well as a number of fixes and code
1020 cleanup (among others, algorithm keywords are now sorted
1021 lexicographically to avoid constant rewrites).
1024 *) In copy_email() check for >= 0 as a return value for
1025 X509_NAME_get_index_by_NID() since 0 is a valid index.
1026 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
1028 +) In BN_div() keep a copy of the sign of 'num' before writing the
1029 result to 'rm' because if rm==num the value will be overwritten
1030 and produce the wrong result if 'num' is negative: this caused
1031 problems with BN_mod() and BN_nnmod().
1034 +) Function OCSP_request_verify(). This checks the signature on an
1035 OCSP request and verifies the signer certificate. The signer
1036 certificate is just checked for a generic purpose and OCSP request
1040 +) Add OCSP_check_validity() function to check the validity of OCSP
1041 responses. OCSP responses are prepared in real time and may only
1042 be a few seconds old. Simply checking that the current time lies
1043 between thisUpdate and nextUpdate max reject otherwise valid responses
1044 caused by either OCSP responder or client clock inaccuracy. Instead
1045 we allow thisUpdate and nextUpdate to fall within a certain period of
1046 the current time. The age of the response can also optionally be
1047 checked. Two new options -validity_period and -status_age added to
1051 +) If signature or public key algorithm is unrecognized print out its
1052 OID rather that just UNKNOWN.
1055 *) Avoid coredump with unsupported or invalid public keys by checking if
1056 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
1057 PKCS7_verify() fails with non detached data.
1060 +) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
1061 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
1062 ID to be generated from the issuer certificate alone which can then be
1063 passed to OCSP_id_issuer_cmp().
1066 +) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
1067 ASN1 modules to export functions returning ASN1_ITEM pointers
1068 instead of the ASN1_ITEM structures themselves. This adds several
1069 new macros which allow the underlying ASN1 function/structure to
1070 be accessed transparently. As a result code should not use ASN1_ITEM
1071 references directly (such as &X509_it) but instead use the relevant
1072 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
1073 use of the new ASN1 code on platforms where exporting structures
1074 is problematical (for example in shared libraries) but exporting
1075 functions returning pointers to structures is not.
1078 +) Add support for overriding the generation of SSL/TLS session IDs.
1079 These callbacks can be registered either in an SSL_CTX or per SSL.
1080 The purpose of this is to allow applications to control, if they wish,
1081 the arbitrary values chosen for use as session IDs, particularly as it
1082 can be useful for session caching in multiple-server environments. A
1083 command-line switch for testing this (and any client code that wishes
1084 to use such a feature) has been added to "s_server".
1085 [Geoff Thorpe, Lutz Jaenicke]
1087 +) Modify mkdef.pl to recognise and parse preprocessor conditionals
1088 of the form '#if defined(...) || defined(...) || ...' and
1089 '#if !defined(...) && !defined(...) && ...'. This also avoids
1090 the growing number of special cases it was previously handling.
1093 *) Don't use getenv in library functions when run as setuid/setgid.
1094 New function OPENSSL_issetugid().
1097 +) Make all configuration macros available for application by making
1098 sure they are available in opensslconf.h, by giving them names starting
1099 with "OPENSSL_" to avoid conflicts with other packages and by making
1100 sure e_os2.h will cover all platform-specific cases together with
1102 Additionally, it is now possible to define configuration/platform-
1103 specific names (called "system identities"). In the C code, these
1104 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
1105 macro with the name beginning with "OPENSSL_SYS_", which is determined
1106 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
1110 +) New option -set_serial to 'req' and 'x509' this allows the serial
1111 number to use to be specified on the command line. Previously self
1112 signed certificates were hard coded with serial number 0 and the
1113 CA options of 'x509' had to use a serial number in a file which was
1117 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
1118 due to incorrect handling of multi-threading:
1120 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
1122 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
1124 3. Count how many times MemCheck_off() has been called so that
1125 nested use can be treated correctly. This also avoids
1126 inband-signalling in the previous code (which relied on the
1127 assumption that thread ID 0 is impossible).
1130 +) New options to 'ca' utility to support V2 CRL entry extensions.
1131 Currently CRL reason, invalidity date and hold instruction are
1132 supported. Add new CRL extensions to V3 code and some new objects.
1135 *) Add "-rand" option also to s_client and s_server.
1138 +) New function EVP_CIPHER_CTX_set_padding() this is used to
1139 disable standard block padding (aka PKCS#5 padding) in the EVP
1140 API, which was previously mandatory. This means that the data is
1141 not padded in any way and so the total length much be a multiple
1142 of the block size, otherwise an error occurs.
1145 +) Initial (incomplete) OCSP SSL support.
1148 *) Fix CPU detection on Irix 6.x.
1149 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
1150 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
1152 +) New function OCSP_parse_url(). This splits up a URL into its host,
1153 port and path components: primarily to parse OCSP URLs. New -url
1154 option to ocsp utility.
1157 +) New nonce behavior. The return value of OCSP_check_nonce() now
1158 reflects the various checks performed. Applications can decide
1159 whether to tolerate certain situations such as an absent nonce
1160 in a response when one was present in a request: the ocsp application
1161 just prints out a warning. New function OCSP_add1_basic_nonce()
1162 this is to allow responders to include a nonce in a response even if
1163 the request is nonce-less.
1166 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
1169 [This change does not apply to 0.9.7.]
1171 *) Use the cached encoding of an X509_NAME structure rather than
1172 copying it. This is apparently the reason for the libsafe "errors"
1173 but the code is actually correct.
1176 +) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
1177 skipped when using openssl x509 multiple times on a single input file,
1178 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
1181 +) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
1182 set string type: to handle setting ASN1_TIME structures. Fix ca
1183 utility to correctly initialize revocation date of CRLs.
1186 +) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
1187 the clients preferred ciphersuites and rather use its own preferences.
1188 Should help to work around M$ SGC (Server Gated Cryptography) bug in
1189 Internet Explorer by ensuring unchanged hash method during stepup.
1190 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
1193 +) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
1194 to aes and add a new 'exist' option to print out symbols that don't
1198 +) Additional options to ocsp utility to allow flags to be set and
1199 additional certificates supplied.
1202 +) Add the option -VAfile to 'openssl ocsp', so the user can give the
1203 OCSP client a number of certificate to only verify the response
1207 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
1208 Bleichenbacher's DSA attack.
1209 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
1210 to be set and top=0 forces the highest bit to be set; top=-1 is new
1211 and leaves the highest bit random.
1212 [Ulf Moeller, Bodo Moeller]
1214 +) Update Rijndael code to version 3.0 and change EVP AES ciphers to
1215 handle the new API. Currently only ECB, CBC modes supported. Add new
1216 AES OIDs. Add TLS AES ciphersuites as described in the "AES Ciphersuites
1217 for TLS" draft-ietf-tls-ciphersuite-03.txt.
1218 [Ben Laurie, Steve Henson]
1220 *) In the NCONF_...-based implementations for CONF_... queries
1221 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
1222 a temporary CONF structure with the data component set to NULL
1223 (which gives segmentation faults in lh_retrieve).
1224 Instead, use NULL for the CONF pointer in CONF_get_string and
1225 CONF_get_number (which may use environment variables) and directly
1226 return NULL from CONF_get_section.
1229 *) Fix potential buffer overrun for EBCDIC.
1232 +) New function OCSP_copy_nonce() to copy nonce value (if present) from
1233 request to response.
1236 +) Functions for OCSP responders. OCSP_request_onereq_count(),
1237 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
1238 extract information from a certificate request. OCSP_response_create()
1239 creates a response and optionally adds a basic response structure.
1240 OCSP_basic_add1_status() adds a complete single response to a basic
1241 response and returns the OCSP_SINGLERESP structure just added (to allow
1242 extensions to be included for example). OCSP_basic_add1_cert() adds a
1243 certificate to a basic response and OCSP_basic_sign() signs a basic
1244 response with various flags. New helper functions ASN1_TIME_check()
1245 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
1246 (converts ASN1_TIME to GeneralizedTime).
1249 +) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
1250 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
1251 structure from a certificate. X509_pubkey_digest() digests the public_key
1252 contents: this is used in various key identifiers.
1255 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
1256 keyUsage if basicConstraints absent for a CA.
1259 *) Make SMIME_write_PKCS7() write mail header values with a format that
1260 is more generally accepted (no spaces before the semicolon), since
1261 some programs can't parse those values properly otherwise. Also make
1262 sure BIO's that break lines after each write do not create invalid
1266 +) Make sk_sort() tolerate a NULL argument.
1267 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
1269 +) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
1270 passed by the function are trusted implicitly. If any of them signed the
1271 response then it is assumed to be valid and is not verified.
1274 -) Make the CRL encoding routines work with empty SEQUENCE OF. The
1275 macros previously used would not encode an empty SEQUENCE OF
1276 and break the signature.
1278 [This change does not apply to 0.9.7.]
1280 *) Zero the premaster secret after deriving the master secret in
1284 +) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
1285 to data. This was previously part of the PKCS7 ASN1 code. This
1286 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
1287 [Steve Henson, reported by Kenneth R. Robinette
1288 <support@securenetterm.com>]
1290 +) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
1291 routines: without these tracing memory leaks is very painful.
1292 Fix leaks in PKCS12 and PKCS7 routines.
1295 *) Add some EVP_add_digest_alias registrations (as found in
1296 OpenSSL_add_all_digests()) to SSL_library_init()
1297 aka OpenSSL_add_ssl_algorithms(). This provides improved
1298 compatibility with peers using X.509 certificates
1299 with unconventional AlgorithmIdentifier OIDs.
1302 *) Fix for Irix with NO_ASM.
1303 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
1305 *) ./config script fixes.
1306 [Ulf Moeller, Richard Levitte]
1308 +) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
1309 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
1310 effectively meant GeneralizedTime would never be used. Now it
1311 is initialised to -1 but X509_time_adj() now has to check the value
1312 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
1313 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
1314 [Steve Henson, reported by Kenneth R. Robinette
1315 <support@securenetterm.com>]
1317 +) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
1318 result in a zero length in the ASN1_INTEGER structure which was
1319 not consistent with the structure when d2i_ASN1_INTEGER() was used
1320 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
1321 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
1322 where it did not print out a minus for negative ASN1_INTEGER.
1325 *) Fix 'openssl passwd -1'.
1328 +) Add summary printout to ocsp utility. The various functions which
1329 convert status values to strings have been renamed to:
1330 OCSP_response_status_str(), OCSP_cert_status_str() and
1331 OCSP_crl_reason_str() and are no longer static. New options
1332 to verify nonce values and to disable verification. OCSP response
1333 printout format cleaned up.
1336 +) Add additional OCSP certificate checks. These are those specified
1337 in RFC2560. This consists of two separate checks: the CA of the
1338 certificate being checked must either be the OCSP signer certificate
1339 or the issuer of the OCSP signer certificate. In the latter case the
1340 OCSP signer certificate must contain the OCSP signing extended key
1341 usage. This check is performed by attempting to match the OCSP
1342 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
1343 in the OCSP_CERTID structures of the response.
1346 +) Initial OCSP certificate verification added to OCSP_basic_verify()
1347 and related routines. This uses the standard OpenSSL certificate
1348 verify routines to perform initial checks (just CA validity) and
1349 to obtain the certificate chain. Then additional checks will be
1350 performed on the chain. Currently the root CA is checked to see
1351 if it is explicitly trusted for OCSP signing. This is used to set
1352 a root CA as a global signing root: that is any certificate that
1353 chains to that CA is an acceptable OCSP signing certificate.
1356 +) New '-extfile ...' option to 'openssl ca' for reading X.509v3
1357 extensions from a separate configuration file.
1358 As when reading extensions from the main configuration file,
1359 the '-extensions ...' option may be used for specifying the
1361 [Massimiliano Pala <madwolf@comune.modena.it>]
1363 *) Change PKCS12_key_gen_asc() so it can cope with non null
1364 terminated strings whose length is passed in the passlen
1365 parameter, for example from PEM callbacks. This was done
1366 by adding an extra length parameter to asc2uni().
1367 [Steve Henson, reported by <oddissey@samsung.co.kr>]
1369 +) New OCSP utility. Allows OCSP requests to be generated or
1370 read. The request can be sent to a responder and the output
1371 parsed, outputed or printed in text form. Not complete yet:
1372 still needs to check the OCSP response validity.
1375 +) New subcommands for 'openssl ca':
1376 'openssl ca -status <serial>' prints the status of the cert with
1377 the given serial number (according to the index file).
1378 'openssl ca -updatedb' updates the expiry status of certificates
1380 [Massimiliano Pala <madwolf@comune.modena.it>]
1382 +) New '-newreq-nodes' command option to CA.pl. This is like
1383 '-newreq', but calls 'openssl req' with the '-nodes' option
1384 so that the resulting key is not encrypted.
1385 [Damien Miller <djm@mindrot.org>]
1387 +) New configuration for the GNU Hurd.
1388 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
1390 +) Initial code to implement OCSP basic response verify. This
1391 is currently incomplete. Currently just finds the signer's
1392 certificate and verifies the signature on the response.
1395 +) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
1396 value of OPENSSLDIR. This is available via the new '-d' option
1397 to 'openssl version', and is also included in 'openssl version -a'.
1400 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
1401 call failed, free the DSA structure.
1404 +) Allowing defining memory allocation callbacks that will be given
1405 file name and line number information in additional arguments
1406 (a const char* and an int). The basic functionality remains, as
1407 well as the original possibility to just replace malloc(),
1408 realloc() and free() by functions that do not know about these
1409 additional arguments. To register and find out the current
1410 settings for extended allocation functions, the following
1411 functions are provided:
1413 CRYPTO_set_mem_ex_functions
1414 CRYPTO_set_locked_mem_ex_functions
1415 CRYPTO_get_mem_ex_functions
1416 CRYPTO_get_locked_mem_ex_functions
1418 These work the same way as CRYPTO_set_mem_functions and friends.
1419 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
1420 extended allocation function is enabled.
1421 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
1422 a conventional allocation function is enabled.
1423 [Richard Levitte, Bodo Moeller]
1425 *) Fix to uni2asc() to cope with zero length Unicode strings.
1426 These are present in some PKCS#12 files.
1429 +) Finish off removing the remaining LHASH function pointer casts.
1430 There should no longer be any prototype-casting required when using
1431 the LHASH abstraction, and any casts that remain are "bugs". See
1432 the callback types and macros at the head of lhash.h for details
1433 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
1436 +) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
1437 If /dev/[u]random devices are not available or do not return enough
1438 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
1440 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
1441 /etc/entropy will be queried once each in this sequence, quering stops
1442 when enough entropy was collected without querying more sockets.
1445 +) Change the Unix RAND_poll() variant to be able to poll several
1446 random devices, as specified by DEVRANDOM, until a sufficient amount
1447 of data has been collected. We spend at most 10 ms on each file
1448 (select timeout) and read in non-blocking mode. DEVRANDOM now
1449 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
1450 (previously it was just the string "/dev/urandom"), so on typical
1451 platforms the 10 ms delay will never occur.
1452 Also separate out the Unix variant to its own file, rand_unix.c.
1453 For VMS, there's a currently-empty rand_vms.c.
1456 +) Move OCSP client related routines to ocsp_cl.c. These
1457 provide utility functions which an application needing
1458 to issue a request to an OCSP responder and analyse the
1459 response will typically need: as opposed to those which an
1460 OCSP responder itself would need which will be added later.
1462 OCSP_request_sign() signs an OCSP request with an API similar
1463 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
1464 response. OCSP_response_get1_basic() extracts basic response
1465 from response. OCSP_resp_find_status(): finds and extracts status
1466 information from an OCSP_CERTID structure (which will be created
1467 when the request structure is built). These are built from lower
1468 level functions which work on OCSP_SINGLERESP structures but
1469 wont normally be used unless the application wishes to examine
1470 extensions in the OCSP response for example.
1472 Replace nonce routines with a pair of functions.
1473 OCSP_request_add1_nonce() adds a nonce value and optionally
1474 generates a random value. OCSP_check_nonce() checks the
1475 validity of the nonce in an OCSP response.
1478 +) Change function OCSP_request_add() to OCSP_request_add0_id().
1479 This doesn't copy the supplied OCSP_CERTID and avoids the
1480 need to free up the newly created id. Change return type
1481 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
1482 This can then be used to add extensions to the request.
1483 Deleted OCSP_request_new(), since most of its functionality
1484 is now in OCSP_REQUEST_new() (and the case insensitive name
1485 clash) apart from the ability to set the request name which
1486 will be added elsewhere.
1489 +) Update OCSP API. Remove obsolete extensions argument from
1490 various functions. Extensions are now handled using the new
1491 OCSP extension code. New simple OCSP HTTP function which
1492 can be used to send requests and parse the response.
1495 +) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
1496 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
1497 uses the special reorder version of SET OF to sort the attributes
1498 and reorder them to match the encoded order. This resolves a long
1499 standing problem: a verify on a PKCS7 structure just after signing
1500 it used to fail because the attribute order did not match the
1501 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
1502 it uses the received order. This is necessary to tolerate some broken
1503 software that does not order SET OF. This is handled by encoding
1504 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
1505 to produce the required SET OF.
1508 +) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
1509 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
1510 files to get correct declarations of the ASN.1 item variables.
1513 +) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
1514 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
1515 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
1516 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
1517 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
1518 ASN1_ITEM and no wrapper functions.
1521 +) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
1522 replace the old function pointer based I/O routines. Change most of
1523 the *_d2i_bio() and *_d2i_fp() functions to use these.
1526 +) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
1527 lines, recognice more "algorithms" that can be deselected, and make
1528 it complain about algorithm deselection that isn't recognised.
1531 +) New ASN1 functions to handle dup, sign, verify, digest, pack and
1532 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
1533 to use new functions. Add NO_ASN1_OLD which can be set to remove
1534 some old style ASN1 functions: this can be used to determine if old
1535 code will still work when these eventually go away.
1538 +) New extension functions for OCSP structures, these follow the
1539 same conventions as certificates and CRLs.
1542 +) New function X509V3_add1_i2d(). This automatically encodes and
1543 adds an extension. Its behaviour can be customised with various
1544 flags to append, replace or delete. Various wrappers added for
1545 certifcates and CRLs.
1548 +) Fix to avoid calling the underlying ASN1 print routine when
1549 an extension cannot be parsed. Correct a typo in the
1550 OCSP_SERVICELOC extension. Tidy up print OCSP format.
1553 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
1554 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
1555 when writing a 32767 byte record.
1556 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
1558 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
1559 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
1561 (RSA objects have a reference count access to which is protected
1562 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
1563 so they are meant to be shared between threads.)
1564 [Bodo Moeller, Geoff Thorpe; original patch submitted by
1565 "Reddie, Steven" <Steven.Reddie@ca.com>]
1567 +) Make mkdef.pl parse some of the ASN1 macros and add apropriate
1568 entries for variables.
1571 *) Fix a deadlock in CRYPTO_mem_leaks().
1574 +) Add functionality to apps/openssl.c for detecting locking
1575 problems: As the program is single-threaded, all we have
1576 to do is register a locking callback using an array for
1577 storing which locks are currently held by the program.
1580 +) Use a lock around the call to CRYPTO_get_ex_new_index() in
1581 SSL_get_ex_data_X509_STORE_idx(), which is used in
1582 ssl_verify_cert_chain() and thus can be called at any time
1583 during TLS/SSL handshakes so that thread-safety is essential.
1584 Unfortunately, the ex_data design is not at all suited
1585 for multi-threaded use, so it probably should be abolished.
1588 +) Added Broadcom "ubsec" ENGINE to OpenSSL.
1589 [Broadcom, tweaked and integrated by Geoff Thorpe]
1591 +) Move common extension printing code to new function
1592 X509V3_print_extensions(). Reorganise OCSP print routines and
1593 implement some needed OCSP ASN1 functions. Add OCSP extensions.
1596 +) New function X509_signature_print() to remove duplication in some
1600 +) Add a special meaning when SET OF and SEQUENCE OF flags are both
1601 set (this was treated exactly the same as SET OF previously). This
1602 is used to reorder the STACK representing the structure to match the
1603 encoding. This will be used to get round a problem where a PKCS7
1604 structure which was signed could not be verified because the STACK
1605 order did not reflect the encoded order.
1608 +) Reimplement the OCSP ASN1 module using the new code.
1611 +) Update the X509V3 code to permit the use of an ASN1_ITEM structure
1612 for its ASN1 operations. The old style function pointers still exist
1613 for now but they will eventually go away.
1616 +) Merge in replacement ASN1 code from the ASN1 branch. This almost
1617 completely replaces the old ASN1 functionality with a table driven
1618 encoder and decoder which interprets an ASN1_ITEM structure describing
1619 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
1620 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
1621 has also been converted to the new form.
1624 +) Change BN_mod_exp_recp so that negative moduli are tolerated
1625 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
1626 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
1627 for negative moduli.
1630 +) Fix BN_uadd and BN_usub: Always return non-negative results instead
1631 of not touching the result's sign bit.
1634 +) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
1638 +) Changed the LHASH code to use prototypes for callbacks, and created
1639 macros to declare and implement thin (optionally static) functions
1640 that provide type-safety and avoid function pointer casting for the
1641 type-specific callbacks.
1644 *) Use better test patterns in bntest.
1647 +) Added Kerberos Cipher Suites to be used with TLS, as written in
1649 [Veers Staats <staatsvr@asc.hpc.mil>,
1650 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
1652 *) rand_win.c fix for Borland C.
1655 *) BN_rshift bugfix for n == 0.
1658 +) Reformat the FAQ so the different questions and answers can be divided
1659 in sections depending on the subject.
1662 +) Have the zlib compression code load ZLIB.DLL dynamically under
1666 +) New function BN_mod_sqrt for computing square roots modulo a prime
1667 (using the probabilistic Tonelli-Shanks algorithm unless
1668 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
1669 be handled deterministically).
1670 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
1672 *) Add a 'bctest' script that checks for some known 'bc' bugs
1673 so that 'make test' does not abort just because 'bc' is broken.
1676 *) Store verify_result within SSL_SESSION also for client side to
1677 avoid potential security hole. (Re-used sessions on the client side
1678 always resulted in verify_result==X509_V_OK, not using the original
1679 result of the server certificate verification.)
1682 +) Make BN_mod_inverse faster by explicitly handling small quotients
1683 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
1684 512 bits], about 30% for larger ones [1024 or 2048 bits].)
1687 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
1688 SSL3_RT_APPLICATION_DATA, return 0.
1689 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
1693 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
1694 releases, have been re-implemented by renaming the previous
1695 implementations of ssl2_read and ssl3_read to ssl2_read_internal
1696 and ssl3_read_internal, respectively, and adding 'peek' parameters
1697 to them. The new ssl[23]_{read,peek} functions are calls to
1698 ssl[23]_read_internal with the 'peek' flag set appropriately.
1699 A 'peek' parameter has also been added to ssl3_read_bytes, which
1700 does the actual work for ssl3_read_internal.
1703 +) New function BN_kronecker.
1706 +) Fix BN_gcd so that it works on negative inputs; the result is
1707 positive unless both parameters are zero.
1708 Previously something reasonably close to an infinite loop was
1709 possible because numbers could be growing instead of shrinking
1710 in the implementation of Euclid's algorithm.
1713 +) Fix BN_is_word() and BN_is_one() macros to take into account the
1714 sign of the number in question.
1716 Fix BN_is_word(a,w) to work correctly for w == 0.
1718 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
1719 because its test if the absolute value of 'a' equals 'w'.
1720 Note that BN_abs_is_word does *not* handle w == 0 reliably;
1721 it exists mostly for use in the implementations of BN_is_zero(),
1722 BN_is_one(), and BN_is_word().
1725 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
1726 the method-specific "init()" handler. Also clean up ex_data after
1727 calling the method-specific "finish()" handler. Previously, this was
1728 happening the other way round.
1731 +) New function BN_swap.
1734 +) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
1735 the exponentiation functions are more likely to produce reasonable
1736 results on negative inputs.
1739 +) Change BN_mod_mul so that the result is always non-negative.
1740 Previously, it could be negative if one of the factors was negative;
1741 I don't think anyone really wanted that behaviour.
1744 +) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
1745 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
1746 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
1747 and add new functions:
1756 BN_mod_lshift1_quick
1760 These functions always generate non-negative results.
1762 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
1763 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
1765 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
1766 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
1767 be reduced modulo m.
1768 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
1770 +) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
1771 was actually never needed) and in BN_mul(). The removal in BN_mul()
1772 required a small change in bn_mul_part_recursive() and the addition
1773 of the functions bn_cmp_part_words(), bn_sub_part_words() and
1774 bn_add_part_words(), which do the same thing as bn_cmp_words(),
1775 bn_sub_words() and bn_add_words() except they take arrays with
1779 +) In 'openssl passwd', verify passwords read from the terminal
1780 unless the '-salt' option is used (which usually means that
1781 verification would just waste user's time since the resulting
1782 hash is going to be compared with some given password hash)
1783 or the new '-noverify' option is used.
1785 This is an incompatible change, but it does not affect
1786 non-interactive use of 'openssl passwd' (passwords on the command
1787 line, '-stdin' option, '-in ...' option) and thus should not
1791 +) Remove all references to RSAref, since there's no more need for it.
1794 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
1795 The previous value, 12, was not always sufficient for BN_mod_exp().
1798 +) Make DSO load along a path given through an environment variable
1799 (SHLIB_PATH) with shl_load().
1802 +) Constify the ENGINE code as a result of BIGNUM constification.
1803 Also constify the RSA code and most things related to it. In a
1804 few places, most notable in the depth of the ASN.1 code, ugly
1805 casts back to non-const were required (to be solved at a later
1809 +) Make it so the openssl application has all engines loaded by default.
1812 +) Constify the BIGNUM routines a little more.
1815 *) Make sure that shared libraries get the internal name engine with
1816 the full version number and not just 0. This should mark the
1817 shared libraries as not backward compatible. Of course, this should
1818 be changed again when we can guarantee backward binary compatibility.
1821 +) Add the following functions:
1823 ENGINE_load_cswift()
1825 ENGINE_load_atalla()
1827 ENGINE_load_builtin_engines()
1829 That way, an application can itself choose if external engines that
1830 are built-in in OpenSSL shall ever be used or not. The benefit is
1831 that applications won't have to be linked with libdl or other dso
1832 libraries unless it's really needed.
1834 Changed 'openssl engine' to load all engines on demand.
1835 Changed the engine header files to avoid the duplication of some
1836 declarations (they differed!).
1839 +) 'openssl engine' can now list capabilities.
1842 +) Better error reporting in 'openssl engine'.
1845 +) Never call load_dh_param(NULL) in s_server.
1848 +) Add engine application. It can currently list engines by name and
1849 identity, and test if they are actually available.
1852 +) Add support for shared libraries under Irix.
1853 [Albert Chin-A-Young <china@thewrittenword.com>]
1855 +) Improve RPM specification file by forcing symbolic linking and making
1856 sure the installed documentation is also owned by root.root.
1857 [Damien Miller <djm@mindrot.org>]
1859 +) Add configuration option to build on Linux on both big-endian and
1861 [Ralf Baechle <ralf@uni-koblenz.de>]
1863 +) Give the OpenSSL applications more possibilities to make use of
1864 keys (public as well as private) handled by engines.
1867 +) Add OCSP code that comes from CertCo.
1870 +) Add VMS support for the Rijndael code.
1873 +) Added untested support for Nuron crypto accelerator.
1876 +) Add support for external cryptographic devices. This code was
1877 previously distributed separately as the "engine" branch.
1878 [Geoff Thorpe, Richard Levitte]
1880 +) Rework the filename-translation in the DSO code. It is now possible to
1881 have far greater control over how a "name" is turned into a filename
1882 depending on the operating environment and any oddities about the
1883 different shared library filenames on each system.
1886 +) Support threads on FreeBSD-elf in Configure.
1889 +) Add the possibility to create shared libraries on HP-UX
1892 +) Fix for SHA1 assembly problem with MASM: it produces
1893 warnings about corrupt line number information when assembling
1894 with debugging information. This is caused by the overlapping
1896 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
1899 NCONF_get_number() has no error checking at all. As a replacement,
1900 NCONF_get_number_e() is defined (_e for "error checking") and is
1901 promoted strongly. The old NCONF_get_number is kept around for
1902 binary backward compatibility.
1903 Make it possible for methods to load from something other than a BIO,
1904 by providing a function pointer that is given a name instead of a BIO.
1905 For example, this could be used to load configuration data from an
1909 *) Fix typo in get_cert_by_subject() in by_dir.c
1910 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
1912 *) Rework the system to generate shared libraries:
1914 - Make note of the expected extension for the shared libraries and
1915 if there is a need for symbolic links from for example libcrypto.so.0
1916 to libcrypto.so.0.9.7. There is extended info in Configure for
1919 - Make as few rebuilds of the shared libraries as possible.
1921 - Still avoid linking the OpenSSL programs with the shared libraries.
1923 - When installing, install the shared libraries separately from the
1927 +) Fix for non blocking accept BIOs. Added new I/O special reason
1928 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
1929 with non blocking I/O was not possible because no retry code was
1930 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
1934 +) Added the beginnings of Rijndael support.
1937 +) Fix for bug in DirectoryString mask setting. Add support for
1938 X509_NAME_print_ex() in 'req' and X509_print_ex() function
1939 to allow certificate printing to more controllable, additional
1940 'certopt' option to 'x509' to allow new printing options to be
1944 +) Clean old EAY MD5 hack from e_os.h.
1947 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
1949 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
1950 and not in SSL_clear because the latter is also used by the
1951 accept/connect functions; previously, the settings made by
1952 SSL_set_read_ahead would be lost during the handshake.
1953 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
1955 *) Correct util/mkdef.pl to be selective about disabled algorithms.
1956 Previously, it would create entries for disableed algorithms no
1960 *) Added several new manual pages for SSL_* function.
1963 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
1965 *) In ssl23_get_client_hello, generate an error message when faced
1966 with an initial SSL 3.0/TLS record that is too small to contain the
1967 first two bytes of the ClientHello message, i.e. client_version.
1968 (Note that this is a pathologic case that probably has never happened
1969 in real life.) The previous approach was to use the version number
1970 from the record header as a substitute; but our protocol choice
1971 should not depend on that one because it is not authenticated
1972 by the Finished messages.
1975 *) More robust randomness gathering functions for Windows.
1976 [Jeffrey Altman <jaltman@columbia.edu>]
1978 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
1979 not set then we don't setup the error code for issuer check errors
1980 to avoid possibly overwriting other errors which the callback does
1981 handle. If an application does set the flag then we assume it knows
1982 what it is doing and can handle the new informational codes
1986 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
1987 a general "ANY" type, as such it should be able to decode anything
1988 including tagged types. However it didn't check the class so it would
1989 wrongly interpret tagged types in the same way as their universal
1990 counterpart and unknown types were just rejected. Changed so that the
1991 tagged and unknown types are handled in the same way as a SEQUENCE:
1992 that is the encoding is stored intact. There is also a new type
1993 "V_ASN1_OTHER" which is used when the class is not universal, in this
1994 case we have no idea what the actual type is so we just lump them all
1998 *) On VMS, stdout may very well lead to a file that is written to
1999 in a record-oriented fashion. That means that every write() will
2000 write a separate record, which will be read separately by the
2001 programs trying to read from it. This can be very confusing.
2003 The solution is to put a BIO filter in the way that will buffer
2004 text until a linefeed is reached, and then write everything a
2005 line at a time, so every record written will be an actual line,
2006 not chunks of lines and not (usually doesn't happen, but I've
2007 seen it once) several lines in one record. BIO_f_linebuffer() is
2010 Currently, it's a VMS-only method, because that's where it has
2011 been tested well enough.
2014 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
2015 it can return incorrect results.
2016 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
2017 but it was in 0.9.6-beta[12].)
2020 *) Disable the check for content being present when verifying detached
2021 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
2022 include zero length content when signing messages.
2025 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
2026 BIO_ctrl (for BIO pairs).
2029 *) Add DSO method for VMS.
2032 *) Bug fix: Montgomery multiplication could produce results with the
2036 *) Add RPM specification openssl.spec and modify it to build three
2037 packages. The default package contains applications, application
2038 documentation and run-time libraries. The devel package contains
2039 include files, static libraries and function documentation. The
2040 doc package contains the contents of the doc directory. The original
2041 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
2044 *) Add a large number of documentation files for many SSL routines.
2045 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
2047 *) Add a configuration entry for Sony News 4.
2048 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
2050 *) Don't set the two most significant bits to one when generating a
2051 random number < q in the DSA library.
2054 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
2055 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
2056 the underlying transport is blocking) if a handshake took place.
2057 (The default behaviour is needed by applications such as s_client
2058 and s_server that use select() to determine when to use SSL_read;
2059 but for applications that know in advance when to expect data, it
2060 just makes things more complicated.)
2063 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
2067 *) Add a few more EBCDIC conditionals that make `req' and `x509'
2068 work better on such systems.
2069 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
2071 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
2072 Update PKCS12_parse() so it copies the friendlyName and the
2073 keyid to the certificates aux info.
2076 *) Fix bug in PKCS7_verify() which caused an infinite loop
2077 if there was more than one signature.
2078 [Sven Uszpelkat <su@celocom.de>]
2080 *) Major change in util/mkdef.pl to include extra information
2081 about each symbol, as well as presentig variables as well
2082 as functions. This change means that there's n more need
2083 to rebuild the .num files when some algorithms are excluded.
2086 *) Allow the verify time to be set by an application,
2087 rather than always using the current time.
2090 *) Phase 2 verify code reorganisation. The certificate
2091 verify code now looks up an issuer certificate by a
2092 number of criteria: subject name, authority key id
2093 and key usage. It also verifies self signed certificates
2094 by the same criteria. The main comparison function is
2095 X509_check_issued() which performs these checks.
2097 Lot of changes were necessary in order to support this
2098 without completely rewriting the lookup code.
2100 Authority and subject key identifier are now cached.
2102 The LHASH 'certs' is X509_STORE has now been replaced
2103 by a STACK_OF(X509_OBJECT). This is mainly because an
2104 LHASH can't store or retrieve multiple objects with
2105 the same hash value.
2107 As a result various functions (which were all internal
2108 use only) have changed to handle the new X509_STORE
2109 structure. This will break anything that messed round
2110 with X509_STORE internally.
2112 The functions X509_STORE_add_cert() now checks for an
2113 exact match, rather than just subject name.
2115 The X509_STORE API doesn't directly support the retrieval
2116 of multiple certificates matching a given criteria, however
2117 this can be worked round by performing a lookup first
2118 (which will fill the cache with candidate certificates)
2119 and then examining the cache for matches. This is probably
2120 the best we can do without throwing out X509_LOOKUP
2121 entirely (maybe later...).
2123 The X509_VERIFY_CTX structure has been enhanced considerably.
2125 All certificate lookup operations now go via a get_issuer()
2126 callback. Although this currently uses an X509_STORE it
2127 can be replaced by custom lookups. This is a simple way
2128 to bypass the X509_STORE hackery necessary to make this
2129 work and makes it possible to use more efficient techniques
2130 in future. A very simple version which uses a simple
2131 STACK for its trusted certificate store is also provided
2132 using X509_STORE_CTX_trusted_stack().
2134 The verify_cb() and verify() callbacks now have equivalents
2135 in the X509_STORE_CTX structure.
2137 X509_STORE_CTX also has a 'flags' field which can be used
2138 to customise the verify behaviour.
2141 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
2142 excludes S/MIME capabilities.
2145 *) When a certificate request is read in keep a copy of the
2146 original encoding of the signed data and use it when outputing
2147 again. Signatures then use the original encoding rather than
2148 a decoded, encoded version which may cause problems if the
2149 request is improperly encoded.
2152 *) For consistency with other BIO_puts implementations, call
2153 buffer_write(b, ...) directly in buffer_puts instead of calling
2156 In BIO_puts, increment b->num_write as in BIO_write.
2157 [Peter.Sylvester@EdelWeb.fr]
2159 *) Fix BN_mul_word for the case where the word is 0. (We have to use
2160 BN_zero, we may not return a BIGNUM with an array consisting of
2164 *) Avoid calling abort() from within the library when problems are
2165 detected, except if preprocessor symbols have been defined
2166 (such as REF_CHECK, BN_DEBUG etc.).
2169 *) New openssl application 'rsautl'. This utility can be
2170 used for low level RSA operations. DER public key
2171 BIO/fp routines also added.
2174 *) New Configure entry and patches for compiling on QNX 4.
2175 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
2177 *) A demo state-machine implementation was sponsored by
2178 Nuron (http://www.nuron.com/) and is now available in
2179 demos/state_machine.
2182 *) New options added to the 'dgst' utility for signature
2183 generation and verification.
2186 *) Unrecognized PKCS#7 content types are now handled via a
2187 catch all ASN1_TYPE structure. This allows unsupported
2188 types to be stored as a "blob" and an application can
2189 encode and decode it manually.
2192 *) Fix various signed/unsigned issues to make a_strex.c
2194 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
2196 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
2197 length if passed a buffer. ASN1_INTEGER_to_BN failed
2198 if passed a NULL BN and its argument was negative.
2199 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
2201 *) Modification to PKCS#7 encoding routines to output definite
2202 length encoding. Since currently the whole structures are in
2203 memory there's not real point in using indefinite length
2204 constructed encoding. However if OpenSSL is compiled with
2205 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
2208 *) Added BIO_vprintf() and BIO_vsnprintf().
2211 *) Added more prefixes to parse for in the the strings written
2212 through a logging bio, to cover all the levels that are available
2213 through syslog. The prefixes are now:
2215 PANIC, EMERG, EMR => LOG_EMERG
2216 ALERT, ALR => LOG_ALERT
2217 CRIT, CRI => LOG_CRIT
2218 ERROR, ERR => LOG_ERR
2219 WARNING, WARN, WAR => LOG_WARNING
2220 NOTICE, NOTE, NOT => LOG_NOTICE
2221 INFO, INF => LOG_INFO
2222 DEBUG, DBG => LOG_DEBUG
2224 and as before, if none of those prefixes are present at the
2225 beginning of the string, LOG_ERR is chosen.
2227 On Win32, the LOG_* levels are mapped according to this:
2229 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
2230 LOG_WARNING => EVENTLOG_WARNING_TYPE
2231 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
2235 *) Made it possible to reconfigure with just the configuration
2236 argument "reconf" or "reconfigure". The command line arguments
2237 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
2238 and are retrieved from there when reconfiguring.
2242 [Assar Westerlund <assar@sics.se>, Richard Levitte]
2244 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
2247 *) The obj_dat.pl script was messing up the sorting of object
2248 names. The reason was that it compared the quoted version
2249 of strings as a result "OCSP" > "OCSP Signing" because
2250 " > SPACE. Changed script to store unquoted versions of
2251 names and add quotes on output. It was also omitting some
2252 names from the lookup table if they were given a default
2253 value (that is if SN is missing it is given the same
2254 value as LN and vice versa), these are now added on the
2255 grounds that if an object has a name we should be able to
2256 look it up. Finally added warning output when duplicate
2257 short or long names are found.
2260 *) Changes needed for Tandem NSK.
2261 [Scott Uroff <scott@xypro.com>]
2263 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
2264 RSA_padding_check_SSLv23(), special padding was never detected
2265 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
2266 version rollback attacks was not effective.
2268 In s23_clnt.c, don't use special rollback-attack detection padding
2269 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
2270 client; similarly, in s23_srvr.c, don't do the rollback check if
2271 SSL 2.0 is the only protocol enabled in the server.
2274 *) Make it possible to get hexdumps of unprintable data with 'openssl
2275 asn1parse'. By implication, the functions ASN1_parse_dump() and
2276 BIO_dump_indent() are added.
2279 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
2280 these print out strings and name structures based on various
2281 flags including RFC2253 support and proper handling of
2282 multibyte characters. Added options to the 'x509' utility
2283 to allow the various flags to be set.
2286 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
2287 Also change the functions X509_cmp_current_time() and
2288 X509_gmtime_adj() work with an ASN1_TIME structure,
2289 this will enable certificates using GeneralizedTime in validity
2290 dates to be checked.
2293 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
2294 negative public key encodings) on by default,
2295 NO_NEG_PUBKEY_BUG can be set to disable it.
2298 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
2299 content octets. An i2c_ASN1_OBJECT is unnecessary because
2300 the encoding can be trivially obtained from the structure.
2303 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
2304 not read locks (CRYPTO_r_[un]lock).
2307 *) A first attempt at creating official support for shared
2308 libraries through configuration. I've kept it so the
2309 default is static libraries only, and the OpenSSL programs
2310 are always statically linked for now, but there are
2311 preparations for dynamic linking in place.
2312 This has been tested on Linux and Tru64.
2315 *) Randomness polling function for Win9x, as described in:
2316 Peter Gutmann, Software Generation of Practically Strong
2320 *) Fix so PRNG is seeded in req if using an already existing
2324 *) New options to smime application. -inform and -outform
2325 allow alternative formats for the S/MIME message including
2326 PEM and DER. The -content option allows the content to be
2327 specified separately. This should allow things like Netscape
2328 form signing output easier to verify.
2331 *) Fix the ASN1 encoding of tags using the 'long form'.
2334 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
2335 STRING types. These convert content octets to and from the
2336 underlying type. The actual tag and length octets are
2337 already assumed to have been read in and checked. These
2338 are needed because all other string types have virtually
2339 identical handling apart from the tag. By having versions
2340 of the ASN1 functions that just operate on content octets
2341 IMPLICIT tagging can be handled properly. It also allows
2342 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
2343 and ASN1_INTEGER are identical apart from the tag.
2346 *) Change the handling of OID objects as follows:
2348 - New object identifiers are inserted in objects.txt, following
2349 the syntax given in objects.README.
2350 - objects.pl is used to process obj_mac.num and create a new
2352 - obj_dat.pl is used to create a new obj_dat.h, using the data in
2355 This is currently kind of a hack, and the perl code in objects.pl
2356 isn't very elegant, but it works as I intended. The simplest way
2357 to check that it worked correctly is to look in obj_dat.h and
2358 check the array nid_objs and make sure the objects haven't moved
2359 around (this is important!). Additions are OK, as well as
2360 consistent name changes.
2363 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
2366 *) Addition of the command line parameter '-rand file' to 'openssl req'.
2367 The given file adds to whatever has already been seeded into the
2368 random pool through the RANDFILE configuration file option or
2369 environment variable, or the default random state file.
2372 *) mkstack.pl now sorts each macro group into lexical order.
2373 Previously the output order depended on the order the files
2374 appeared in the directory, resulting in needless rewriting
2378 *) Patches to make OpenSSL compile under Win32 again. Mostly
2379 work arounds for the VC++ problem that it treats func() as
2380 func(void). Also stripped out the parts of mkdef.pl that
2381 added extra typesafe functions: these no longer exist.
2384 *) Reorganisation of the stack code. The macros are now all
2385 collected in safestack.h . Each macro is defined in terms of
2386 a "stack macro" of the form SKM_<name>(type, a, b). The
2387 DEBUG_SAFESTACK is now handled in terms of function casts,
2388 this has the advantage of retaining type safety without the
2389 use of additional functions. If DEBUG_SAFESTACK is not defined
2390 then the non typesafe macros are used instead. Also modified the
2391 mkstack.pl script to handle the new form. Needs testing to see
2392 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
2393 the default if no major problems. Similar behaviour for ASN1_SET_OF
2394 and PKCS12_STACK_OF.
2397 *) When some versions of IIS use the 'NET' form of private key the
2398 key derivation algorithm is different. Normally MD5(password) is
2399 used as a 128 bit RC4 key. In the modified case
2400 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
2401 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
2402 as the old Netscape_RSA functions except they have an additional
2403 'sgckey' parameter which uses the modified algorithm. Also added
2404 an -sgckey command line option to the rsa utility. Thanks to
2405 Adrian Peck <bertie@ncipher.com> for posting details of the modified
2406 algorithm to openssl-dev.
2409 *) The evp_local.h macros were using 'c.##kname' which resulted in
2410 invalid expansion on some systems (SCO 5.0.5 for example).
2411 Corrected to 'c.kname'.
2412 [Phillip Porch <root@theporch.com>]
2414 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
2415 a STACK of email addresses from a certificate or request, these look
2416 in the subject name and the subject alternative name extensions and
2417 omit any duplicate addresses.
2420 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
2421 This makes DSA verification about 2 % faster.
2424 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
2425 (meaning that now 2^5 values will be precomputed, which is only 4 KB
2426 plus overhead for 1024 bit moduli).
2427 This makes exponentiations about 0.5 % faster for 1024 bit
2428 exponents (as measured by "openssl speed rsa2048").
2431 *) Rename memory handling macros to avoid conflicts with other
2433 Malloc => OPENSSL_malloc
2434 Malloc_locked => OPENSSL_malloc_locked
2435 Realloc => OPENSSL_realloc
2436 Free => OPENSSL_free
2439 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
2440 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
2443 *) CygWin32 support.
2444 [John Jarvie <jjarvie@newsguy.com>]
2446 *) The type-safe stack code has been rejigged. It is now only compiled
2447 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
2448 by default all type-specific stack functions are "#define"d back to
2449 standard stack functions. This results in more streamlined output
2450 but retains the type-safety checking possibilities of the original
2454 *) The STACK code has been cleaned up, and certain type declarations
2455 that didn't make a lot of sense have been brought in line. This has
2456 also involved a cleanup of sorts in safestack.h to more correctly
2457 map type-safe stack functions onto their plain stack counterparts.
2458 This work has also resulted in a variety of "const"ifications of
2459 lots of the code, especially "_cmp" operations which should normally
2460 be prototyped with "const" parameters anyway.
2463 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
2464 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
2465 (The PRNG state consists of two parts, the large pool 'state' and 'md',
2466 where all of 'md' is used each time the PRNG is used, but 'state'
2467 is used only indexed by a cyclic counter. As entropy may not be
2468 well distributed from the beginning, 'md' is important as a
2469 chaining variable. However, the output function chains only half
2470 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
2471 all of 'md', and seeding with STATE_SIZE dummy bytes will result
2472 in all of 'state' being rewritten, with the new values depending
2473 on virtually all of 'md'. This overcomes the 80 bit limitation.)
2476 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
2477 the handshake is continued after ssl_verify_cert_chain();
2478 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
2479 can lead to 'unexplainable' connection aborts later.
2480 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
2482 *) Major EVP API cipher revision.
2483 Add hooks for extra EVP features. This allows various cipher
2484 parameters to be set in the EVP interface. Support added for variable
2485 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
2486 setting of RC2 and RC5 parameters.
2488 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
2491 Remove lots of duplicated code from the EVP library. For example *every*
2492 cipher init() function handles the 'iv' in the same way according to the
2493 cipher mode. They also all do nothing if the 'key' parameter is NULL and
2494 for CFB and OFB modes they zero ctx->num.
2496 New functionality allows removal of S/MIME code RC2 hack.
2498 Most of the routines have the same form and so can be declared in terms
2501 By shifting this to the top level EVP_CipherInit() it can be removed from
2502 all individual ciphers. If the cipher wants to handle IVs or keys
2503 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
2506 Change lots of functions like EVP_EncryptUpdate() to now return a
2507 value: although software versions of the algorithms cannot fail
2508 any installed hardware versions can.
2511 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
2512 this option is set, tolerate broken clients that send the negotiated
2513 protocol version number instead of the requested protocol version
2517 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
2518 i.e. non-zero for export ciphersuites, zero otherwise.
2519 Previous versions had this flag inverted, inconsistent with
2520 rsa_tmp_cb (..._TMP_RSA_CB).
2521 [Bodo Moeller; problem reported by Amit Chopra]
2523 *) Add missing DSA library text string. Work around for some IIS
2524 key files with invalid SEQUENCE encoding.
2527 *) Add a document (doc/standards.txt) that list all kinds of standards
2528 and so on that are implemented in OpenSSL.
2531 *) Enhance c_rehash script. Old version would mishandle certificates
2532 with the same subject name hash and wouldn't handle CRLs at all.
2533 Added -fingerprint option to crl utility, to support new c_rehash
2537 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
2540 *) Fix for SSL server purpose checking. Server checking was
2541 rejecting certificates which had extended key usage present
2542 but no ssl client purpose.
2543 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
2545 *) Make PKCS#12 code work with no password. The PKCS#12 spec
2546 is a little unclear about how a blank password is handled.
2547 Since the password in encoded as a BMPString with terminating
2548 double NULL a zero length password would end up as just the
2549 double NULL. However no password at all is different and is
2550 handled differently in the PKCS#12 key generation code. NS
2551 treats a blank password as zero length. MSIE treats it as no
2552 password on export: but it will try both on import. We now do
2553 the same: PKCS12_parse() tries zero length and no password if
2554 the password is set to "" or NULL (NULL is now a valid password:
2555 it wasn't before) as does the pkcs12 application.
2558 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
2559 perror when PEM_read_bio_X509_REQ fails, the error message must
2560 be obtained from the error queue.
2563 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
2564 it in ERR_remove_state if appropriate, and change ERR_get_state
2565 accordingly to avoid race conditions (this is necessary because
2566 thread_hash is no longer constant once set).
2569 *) Bugfix for linux-elf makefile.one.
2572 *) RSA_get_default_method() will now cause a default
2573 RSA_METHOD to be chosen if one doesn't exist already.
2574 Previously this was only set during a call to RSA_new()
2575 or RSA_new_method(NULL) meaning it was possible for
2576 RSA_get_default_method() to return NULL.
2579 *) Added native name translation to the existing DSO code
2580 that will convert (if the flag to do so is set) filenames
2581 that are sufficiently small and have no path information
2582 into a canonical native form. Eg. "blah" converted to
2583 "libblah.so" or "blah.dll" etc.
2586 *) New function ERR_error_string_n(e, buf, len) which is like
2587 ERR_error_string(e, buf), but writes at most 'len' bytes
2588 including the 0 terminator. For ERR_error_string_n, 'buf'
2590 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
2592 *) CONF library reworked to become more general. A new CONF
2593 configuration file reader "class" is implemented as well as a
2594 new functions (NCONF_*, for "New CONF") to handle it. The now
2595 old CONF_* functions are still there, but are reimplemented to
2596 work in terms of the new functions. Also, a set of functions
2597 to handle the internal storage of the configuration data is
2598 provided to make it easier to write new configuration file
2599 reader "classes" (I can definitely see something reading a
2600 configuration file in XML format, for example), called _CONF_*,
2601 or "the configuration storage API"...
2603 The new configuration file reading functions are:
2605 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
2606 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
2608 NCONF_default, NCONF_WIN32
2610 NCONF_dump_fp, NCONF_dump_bio
2612 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
2613 NCONF_new creates a new CONF object. This works in the same way
2614 as other interfaces in OpenSSL, like the BIO interface.
2615 NCONF_dump_* dump the internal storage of the configuration file,
2616 which is useful for debugging. All other functions take the same
2617 arguments as the old CONF_* functions wth the exception of the
2618 first that must be a `CONF *' instead of a `LHASH *'.
2620 To make it easer to use the new classes with the old CONF_* functions,
2621 the function CONF_set_default_method is provided.
2624 *) Add '-tls1' option to 'openssl ciphers', which was already
2625 mentioned in the documentation but had not been implemented.
2626 (This option is not yet really useful because even the additional
2627 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
2630 *) Initial DSO code added into libcrypto for letting OpenSSL (and
2631 OpenSSL-based applications) load shared libraries and bind to
2632 them in a portable way.
2633 [Geoff Thorpe, with contributions from Richard Levitte]
2635 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
2637 *) Make sure _lrotl and _lrotr are only used with MSVC.
2639 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
2640 (the default implementation of RAND_status).
2642 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
2643 to '-clrext' (= clear extensions), as intended and documented.
2644 [Bodo Moeller; inconsistency pointed out by Michael Attili
2647 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
2648 was larger than the MD block size.
2649 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
2651 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
2652 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
2653 using the passed key: if the passed key was a private key the result
2654 of X509_print(), for example, would be to print out all the private key
2658 *) des_quad_cksum() byte order bug fix.
2659 [Ulf Möller, using the problem description in krb4-0.9.7, where
2660 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
2662 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
2664 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
2666 *) For easily testing in shell scripts whether some command
2667 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
2668 returns with exit code 0 iff no command of the given name is available.
2669 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
2670 the output goes to stdout and nothing is printed to stderr.
2671 Additional arguments are always ignored.
2673 Since for each cipher there is a command of the same name,
2674 the 'no-cipher' compilation switches can be tested this way.
2676 ('openssl no-XXX' is not able to detect pseudo-commands such
2677 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
2680 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
2683 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
2684 is set; it will be thrown away anyway because each handshake creates
2686 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
2687 to parameters -- in previous versions (since OpenSSL 0.9.3) the
2688 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
2689 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
2692 *) New s_client option -ign_eof: EOF at stdin is ignored, and
2693 'Q' and 'R' lose their special meanings (quit/renegotiate).
2694 This is part of what -quiet does; unlike -quiet, -ign_eof
2695 does not suppress any output.
2698 *) Add compatibility options to the purpose and trust code. The
2699 purpose X509_PURPOSE_ANY is "any purpose" which automatically
2700 accepts a certificate or CA, this was the previous behaviour,
2701 with all the associated security issues.
2703 X509_TRUST_COMPAT is the old trust behaviour: only and
2704 automatically trust self signed roots in certificate store. A
2705 new trust setting X509_TRUST_DEFAULT is used to specify that
2706 a purpose has no associated trust setting and it should instead
2707 use the value in the default purpose.
2710 *) Fix the PKCS#8 DSA private key code so it decodes keys again
2711 and fix a memory leak.
2714 *) In util/mkerr.pl (which implements 'make errors'), preserve
2715 reason strings from the previous version of the .c file, as
2716 the default to have only downcase letters (and digits) in
2717 automatically generated reasons codes is not always appropriate.
2720 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
2721 using strerror. Previously, ERR_reason_error_string() returned
2722 library names as reason strings for SYSerr; but SYSerr is a special
2723 case where small numbers are errno values, not library numbers.
2726 *) Add '-dsaparam' option to 'openssl dhparam' application. This
2727 converts DSA parameters into DH parameters. (When creating parameters,
2728 DSA_generate_parameters is used.)
2731 *) Include 'length' (recommended exponent length) in C code generated
2732 by 'openssl dhparam -C'.
2735 *) The second argument to set_label in perlasm was already being used
2736 so couldn't be used as a "file scope" flag. Moved to third argument
2740 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
2741 instead of RAND_bytes for encryption IVs and salts.
2744 *) Include RAND_status() into RAND_METHOD instead of implementing
2745 it only for md_rand.c Otherwise replacing the PRNG by calling
2746 RAND_set_rand_method would be impossible.
2749 *) Don't let DSA_generate_key() enter an infinite loop if the random
2750 number generation fails.
2753 *) New 'rand' application for creating pseudo-random output.
2756 *) Added configuration support for Linux/IA64
2757 [Rolf Haberrecker <rolf@suse.de>]
2759 *) Assembler module support for Mingw32.
2762 *) Shared library support for HPUX (in shlib/).
2763 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
2765 *) Shared library support for Solaris gcc.
2766 [Lutz Behnke <behnke@trustcenter.de>]
2768 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
2770 *) PKCS7_encrypt() was adding text MIME headers twice because they
2771 were added manually and by SMIME_crlf_copy().
2774 *) In bntest.c don't call BN_rand with zero bits argument.
2775 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
2777 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
2778 case was implemented. This caused BN_div_recp() to fail occasionally.
2781 *) Add an optional second argument to the set_label() in the perl
2782 assembly language builder. If this argument exists and is set
2783 to 1 it signals that the assembler should use a symbol whose
2784 scope is the entire file, not just the current function. This
2785 is needed with MASM which uses the format label:: for this scope.
2786 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
2788 *) Change the ASN1 types so they are typedefs by default. Before
2789 almost all types were #define'd to ASN1_STRING which was causing
2790 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
2794 *) Change names of new functions to the new get1/get0 naming
2795 convention: After 'get1', the caller owns a reference count
2796 and has to call ..._free; 'get0' returns a pointer to some
2797 data structure without incrementing reference counters.
2798 (Some of the existing 'get' functions increment a reference
2799 counter, some don't.)
2800 Similarly, 'set1' and 'add1' functions increase reference
2801 counters or duplicate objects.
2804 *) Allow for the possibility of temp RSA key generation failure:
2805 the code used to assume it always worked and crashed on failure.
2808 *) Fix potential buffer overrun problem in BIO_printf().
2809 [Ulf Möller, using public domain code by Patrick Powell; problem
2810 pointed out by David Sacerdote <das33@cornell.edu>]
2812 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
2813 RAND_egd() and RAND_status(). In the command line application,
2814 the EGD socket can be specified like a seed file using RANDFILE
2818 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
2819 Some CAs (e.g. Verisign) distribute certificates in this form.
2822 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
2823 list to exclude them. This means that no special compilation option
2824 is needed to use anonymous DH: it just needs to be included in the
2828 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
2829 EVP_MD_type. The old functionality is available in a new macro called
2830 EVP_MD_md(). Change code that uses it and update docs.
2833 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
2834 where the 'void *' argument is replaced by a function pointer argument.
2835 Previously 'void *' was abused to point to functions, which works on
2836 many platforms, but is not correct. As these functions are usually
2837 called by macros defined in OpenSSL header files, most source code
2838 should work without changes.
2841 *) <openssl/opensslconf.h> (which is created by Configure) now contains
2842 sections with information on -D... compiler switches used for
2843 compiling the library so that applications can see them. To enable
2844 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
2845 must be defined. E.g.,
2846 #define OPENSSL_ALGORITHM_DEFINES
2847 #include <openssl/opensslconf.h>
2848 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
2849 [Richard Levitte, Ulf and Bodo Möller]
2851 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
2855 *) Change the 'other' type in certificate aux info to a STACK_OF
2856 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
2857 the required ASN1 format: arbitrary types determined by an OID.
2860 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
2861 argument to 'req'. This is not because the function is newer or
2862 better than others it just uses the work 'NEW' in the certificate
2863 request header lines. Some software needs this.
2866 *) Reorganise password command line arguments: now passwords can be
2867 obtained from various sources. Delete the PEM_cb function and make
2868 it the default behaviour: i.e. if the callback is NULL and the
2869 usrdata argument is not NULL interpret it as a null terminated pass
2870 phrase. If usrdata and the callback are NULL then the pass phrase
2871 is prompted for as usual.
2874 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
2875 the support is automatically enabled. The resulting binaries will
2876 autodetect the card and use it if present.
2877 [Ben Laurie and Compaq Inc.]
2879 *) Work around for Netscape hang bug. This sends certificate request
2880 and server done in one record. Since this is perfectly legal in the
2881 SSL/TLS protocol it isn't a "bug" option and is on by default. See
2882 the bugs/SSLv3 entry for more info.
2885 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
2888 *) Add -rand argument to smime and pkcs12 applications and read/write
2892 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
2895 *) Add command line password options to the remaining applications.
2898 *) Bug fix for BN_div_recp() for numerators with an even number of
2902 *) More tests in bntest.c, and changed test_bn output.
2905 *) ./config recognizes MacOS X now.
2908 *) Bug fix for BN_div() when the first words of num and divsor are
2909 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
2912 *) Add support for various broken PKCS#8 formats, and command line
2913 options to produce them.
2916 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
2917 get temporary BIGNUMs from a BN_CTX.
2920 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
2924 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
2925 include a #define from the old name to the new. The original intent
2926 was that statically linked binaries could for example just call
2927 SSLeay_add_all_ciphers() to just add ciphers to the table and not
2928 link with digests. This never worked becayse SSLeay_add_all_digests()
2929 and SSLeay_add_all_ciphers() were in the same source file so calling
2930 one would link with the other. They are now in separate source files.
2933 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
2936 *) Use a less unusual form of the Miller-Rabin primality test (it used
2937 a binary algorithm for exponentiation integrated into the Miller-Rabin
2938 loop, our standard modexp algorithms are faster).
2941 *) Support for the EBCDIC character set completed.
2942 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
2944 *) Source code cleanups: use const where appropriate, eliminate casts,
2945 use void * instead of char * in lhash.
2948 *) Bugfix: ssl3_send_server_key_exchange was not restartable
2949 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
2950 this the server could overwrite ephemeral keys that the client
2954 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
2955 using 50 iterations of the Rabin-Miller test.
2957 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
2958 iterations of the Rabin-Miller test as required by the appendix
2959 to FIPS PUB 186[-1]) instead of DSA_is_prime.
2960 As BN_is_prime_fasttest includes trial division, DSA parameter
2961 generation becomes much faster.
2963 This implies a change for the callback functions in DSA_is_prime
2964 and DSA_generate_parameters: The callback function is called once
2965 for each positive witness in the Rabin-Miller test, not just
2966 occasionally in the inner loop; and the parameters to the
2967 callback function now provide an iteration count for the outer
2968 loop rather than for the current invocation of the inner loop.
2969 DSA_generate_parameters additionally can call the callback
2970 function with an 'iteration count' of -1, meaning that a
2971 candidate has passed the trial division test (when q is generated
2972 from an application-provided seed, trial division is skipped).
2975 *) New function BN_is_prime_fasttest that optionally does trial
2976 division before starting the Rabin-Miller test and has
2977 an additional BN_CTX * argument (whereas BN_is_prime always
2978 has to allocate at least one BN_CTX).
2979 'callback(1, -1, cb_arg)' is called when a number has passed the
2980 trial division stage.
2983 *) Fix for bug in CRL encoding. The validity dates weren't being handled
2987 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
2990 *) New function BN_pseudo_rand().
2993 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
2994 bignum version of BN_from_montgomery() with the working code from
2995 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
2999 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
3000 made it impossible to use the same SSL_SESSION data structure in
3001 SSL2 clients in multiple threads.
3004 *) The return value of RAND_load_file() no longer counts bytes obtained
3005 by stat(). RAND_load_file(..., -1) is new and uses the complete file
3006 to seed the PRNG (previously an explicit byte count was required).
3007 [Ulf Möller, Bodo Möller]
3009 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
3010 used (char *) instead of (void *) and had casts all over the place.
3013 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
3016 *) Retain source code compatibility for BN_prime_checks macro:
3017 BN_is_prime(..., BN_prime_checks, ...) now uses
3018 BN_prime_checks_for_size to determine the appropriate number of
3019 Rabin-Miller iterations.
3022 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
3023 DH_CHECK_P_NOT_SAFE_PRIME.
3024 (Check if this is true? OpenPGP calls them "strong".)
3027 *) Merge the functionality of "dh" and "gendh" programs into a new program
3028 "dhparam". The old programs are retained for now but will handle DH keys
3029 (instead of parameters) in future.
3032 *) Make the ciphers, s_server and s_client programs check the return values
3033 when a new cipher list is set.
3036 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
3037 ciphers. Before when the 56bit ciphers were enabled the sorting was
3040 The syntax for the cipher sorting has been extended to support sorting by
3041 cipher-strength (using the strength_bits hard coded in the tables).
3042 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
3044 Fix a bug in the cipher-command parser: when supplying a cipher command
3045 string with an "undefined" symbol (neither command nor alphanumeric
3046 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
3047 an error is flagged.
3049 Due to the strength-sorting extension, the code of the
3050 ssl_create_cipher_list() function was completely rearranged. I hope that
3051 the readability was also increased :-)
3052 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
3054 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
3055 for the first serial number and places 2 in the serial number file. This
3056 avoids problems when the root CA is created with serial number zero and
3057 the first user certificate has the same issuer name and serial number
3061 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
3062 the new code. Add documentation for this stuff.
3065 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
3066 X509_*() to X509at_*() on the grounds that they don't handle X509
3067 structures and behave in an analagous way to the X509v3 functions:
3068 they shouldn't be called directly but wrapper functions should be used
3071 So we also now have some wrapper functions that call the X509at functions
3072 when passed certificate requests. (TO DO: similar things can be done with
3073 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
3074 things. Some of these need some d2i or i2d and print functionality
3075 because they handle more complex structures.)
3078 *) Add missing #ifndefs that caused missing symbols when building libssl
3079 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
3080 NO_RSA in ssl/s2*.c.
3081 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
3083 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
3084 has a return value which indicates the quality of the random data
3085 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
3086 error queue. New function RAND_pseudo_bytes() generates output that is
3087 guaranteed to be unique but not unpredictable. RAND_add is like
3088 RAND_seed, but takes an extra argument for an entropy estimate
3089 (RAND_seed always assumes full entropy).
3092 *) Do more iterations of Rabin-Miller probable prime test (specifically,
3093 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
3094 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
3095 in crypto/bn/bn_prime.c for the complete table). This guarantees a
3096 false-positive rate of at most 2^-80 for random input.
3099 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
3102 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
3103 in the 0.9.5 release), this returns the chain
3104 from an X509_CTX structure with a dup of the stack and all
3105 the X509 reference counts upped: so the stack will exist
3106 after X509_CTX_cleanup() has been called. Modify pkcs12.c
3109 Also make SSL_SESSION_print() print out the verify return
3113 *) Add manpage for the pkcs12 command. Also change the default
3114 behaviour so MAC iteration counts are used unless the new
3115 -nomaciter option is used. This improves file security and
3116 only older versions of MSIE (4.0 for example) need it.
3119 *) Honor the no-xxx Configure options when creating .DEF files.
3122 *) Add PKCS#10 attributes to field table: challengePassword,
3123 unstructuredName and unstructuredAddress. These are taken from
3124 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
3125 international characters are used.
3127 More changes to X509_ATTRIBUTE code: allow the setting of types
3128 based on strings. Remove the 'loc' parameter when adding
3129 attributes because these will be a SET OF encoding which is sorted
3133 *) Initial changes to the 'req' utility to allow request generation
3134 automation. This will allow an application to just generate a template
3135 file containing all the field values and have req construct the
3138 Initial support for X509_ATTRIBUTE handling. Stacks of these are
3139 used all over the place including certificate requests and PKCS#7
3140 structures. They are currently handled manually where necessary with
3141 some primitive wrappers for PKCS#7. The new functions behave in a
3142 manner analogous to the X509 extension functions: they allow
3143 attributes to be looked up by NID and added.
3145 Later something similar to the X509V3 code would be desirable to
3146 automatically handle the encoding, decoding and printing of the
3147 more complex types. The string types like challengePassword can
3148 be handled by the string table functions.
3150 Also modified the multi byte string table handling. Now there is
3151 a 'global mask' which masks out certain types. The table itself
3152 can use the flag STABLE_NO_MASK to ignore the mask setting: this
3153 is useful when for example there is only one permissible type
3154 (as in countryName) and using the mask might result in no valid
3158 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
3159 SSL_get_peer_finished to allow applications to obtain the latest
3160 Finished messages sent to the peer or expected from the peer,
3161 respectively. (SSL_get_peer_finished is usually the Finished message
3162 actually received from the peer, otherwise the protocol will be aborted.)
3164 As the Finished message are message digests of the complete handshake
3165 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
3166 be used for external authentication procedures when the authentication
3167 provided by SSL/TLS is not desired or is not enough.
3170 *) Enhanced support for Alpha Linux is added. Now ./config checks if
3171 the host supports BWX extension and if Compaq C is present on the
3172 $PATH. Just exploiting of the BWX extension results in 20-30%
3173 performance kick for some algorithms, e.g. DES and RC4 to mention
3174 a couple. Compaq C in turn generates ~20% faster code for MD5 and
3178 *) Add support for MS "fast SGC". This is arguably a violation of the
3179 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
3180 weak crypto and after checking the certificate is SGC a second one
3181 with strong crypto. MS SGC stops the first handshake after receiving
3182 the server certificate message and sends a second client hello. Since
3183 a server will typically do all the time consuming operations before
3184 expecting any further messages from the client (server key exchange
3185 is the most expensive) there is little difference between the two.
3187 To get OpenSSL to support MS SGC we have to permit a second client
3188 hello message after we have sent server done. In addition we have to
3189 reset the MAC if we do get this second client hello.
3192 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
3193 if a DER encoded private key is RSA or DSA traditional format. Changed
3194 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
3195 format DER encoded private key. Newer code should use PKCS#8 format which
3196 has the key type encoded in the ASN1 structure. Added DER private key
3197 support to pkcs8 application.
3200 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
3201 ciphersuites has been selected (as required by the SSL 3/TLS 1
3202 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
3203 is set, we interpret this as a request to violate the specification
3204 (the worst that can happen is a handshake failure, and 'correct'
3205 behaviour would result in a handshake failure anyway).
3208 *) In SSL_CTX_add_session, take into account that there might be multiple
3209 SSL_SESSION structures with the same session ID (e.g. when two threads
3210 concurrently obtain them from an external cache).
3211 The internal cache can handle only one SSL_SESSION with a given ID,
3212 so if there's a conflict, we now throw out the old one to achieve
3216 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
3217 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
3218 some routines that use cipher OIDs: some ciphers do not have OIDs
3219 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
3223 *) Simplify the trust setting structure and code. Now we just have
3224 two sequences of OIDs for trusted and rejected settings. These will
3225 typically have values the same as the extended key usage extension
3226 and any application specific purposes.
3228 The trust checking code now has a default behaviour: it will just
3229 check for an object with the same NID as the passed id. Functions can
3230 be provided to override either the default behaviour or the behaviour
3231 for a given id. SSL client, server and email already have functions
3232 in place for compatibility: they check the NID and also return "trusted"
3233 if the certificate is self signed.
3236 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
3237 traditional format into an EVP_PKEY structure.
3240 *) Add a password callback function PEM_cb() which either prompts for
3241 a password if usr_data is NULL or otherwise assumes it is a null
3242 terminated password. Allow passwords to be passed on command line
3243 environment or config files in a few more utilities.
3246 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
3247 keys. Add some short names for PKCS#8 PBE algorithms and allow them
3248 to be specified on the command line for the pkcs8 and pkcs12 utilities.
3249 Update documentation.
3252 *) Support for ASN1 "NULL" type. This could be handled before by using
3253 ASN1_TYPE but there wasn't any function that would try to read a NULL
3254 and produce an error if it couldn't. For compatibility we also have
3255 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
3256 don't allocate anything because they don't need to.
3259 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
3261 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
3263 *) Rebuild of the memory allocation routines used by OpenSSL code and
3264 possibly others as well. The purpose is to make an interface that
3265 provide hooks so anyone can build a separate set of allocation and
3266 deallocation routines to be used by OpenSSL, for example memory
3267 pool implementations, or something else, which was previously hard
3268 since Malloc(), Realloc() and Free() were defined as macros having
3269 the values malloc, realloc and free, respectively (except for Win32
3270 compilations). The same is provided for memory debugging code.
3271 OpenSSL already comes with functionality to find memory leaks, but
3272 this gives people a chance to debug other memory problems.
3274 With these changes, a new set of functions and macros have appeared:
3276 CRYPTO_set_mem_debug_functions() [F]
3277 CRYPTO_get_mem_debug_functions() [F]
3278 CRYPTO_dbg_set_options() [F]
3279 CRYPTO_dbg_get_options() [F]
3280 CRYPTO_malloc_debug_init() [M]
3282 The memory debug functions are NULL by default, unless the library
3283 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
3284 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
3285 gives the standard debugging functions that come with OpenSSL) or
3286 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
3287 provided by the library user) must be used. When the standard
3288 debugging functions are used, CRYPTO_dbg_set_options can be used to
3289 request additional information:
3290 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
3291 the CRYPTO_MDEBUG_xxx macro when compiling the library.
3293 Also, things like CRYPTO_set_mem_functions will always give the
3294 expected result (the new set of functions is used for allocation
3295 and deallocation) at all times, regardless of platform and compiler
3298 To finish it up, some functions that were never use in any other
3299 way than through macros have a new API and new semantic:
3302 CRYPTO_dbg_realloc()
3305 All macros of value have retained their old syntax.
3306 [Richard Levitte and Bodo Moeller]
3308 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
3309 ordering of SMIMECapabilities wasn't in "strength order" and there
3310 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
3314 *) Some ASN1 types with illegal zero length encoding (INTEGER,
3315 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
3316 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
3318 *) Merge in my S/MIME library for OpenSSL. This provides a simple
3319 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
3320 functionality to handle multipart/signed properly) and a utility
3321 called 'smime' to call all this stuff. This is based on code I
3322 originally wrote for Celo who have kindly allowed it to be
3323 included in OpenSSL.
3326 *) Add variants des_set_key_checked and des_set_key_unchecked of
3327 des_set_key (aka des_key_sched). Global variable des_check_key
3328 decides which of these is called by des_set_key; this way
3329 des_check_key behaves as it always did, but applications and
3330 the library itself, which was buggy for des_check_key == 1,
3331 have a cleaner way to pick the version they need.
3334 *) New function PKCS12_newpass() which changes the password of a
3338 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
3339 dynamic mix. In both cases the ids can be used as an index into the
3340 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
3341 functions so they accept a list of the field values and the
3342 application doesn't need to directly manipulate the X509_TRUST
3346 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
3350 *) Modify the way the V3 extension code looks up extensions. This now
3351 works in a similar way to the object code: we have some "standard"
3352 extensions in a static table which is searched with OBJ_bsearch()
3353 and the application can add dynamic ones if needed. The file
3354 crypto/x509v3/ext_dat.h now has the info: this file needs to be
3355 updated whenever a new extension is added to the core code and kept
3356 in ext_nid order. There is a simple program 'tabtest.c' which checks
3357 this. New extensions are not added too often so this file can readily
3358 be maintained manually.
3360 There are two big advantages in doing things this way. The extensions
3361 can be looked up immediately and no longer need to be "added" using
3362 X509V3_add_standard_extensions(): this function now does nothing.
3363 [Side note: I get *lots* of email saying the extension code doesn't
3364 work because people forget to call this function]
3365 Also no dynamic allocation is done unless new extensions are added:
3366 so if we don't add custom extensions there is no need to call
3367 X509V3_EXT_cleanup().
3370 *) Modify enc utility's salting as follows: make salting the default. Add a
3371 magic header, so unsalted files fail gracefully instead of just decrypting
3372 to garbage. This is because not salting is a big security hole, so people
3373 should be discouraged from doing it.
3376 *) Fixes and enhancements to the 'x509' utility. It allowed a message
3377 digest to be passed on the command line but it only used this
3378 parameter when signing a certificate. Modified so all relevant
3379 operations are affected by the digest parameter including the
3380 -fingerprint and -x509toreq options. Also -x509toreq choked if a
3381 DSA key was used because it didn't fix the digest.
3384 *) Initial certificate chain verify code. Currently tests the untrusted
3385 certificates for consistency with the verify purpose (which is set
3386 when the X509_STORE_CTX structure is set up) and checks the pathlength.
3388 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
3389 this is because it will reject chains with invalid extensions whereas
3390 every previous version of OpenSSL and SSLeay made no checks at all.
3392 Trust code: checks the root CA for the relevant trust settings. Trust
3393 settings have an initial value consistent with the verify purpose: e.g.
3394 if the verify purpose is for SSL client use it expects the CA to be
3395 trusted for SSL client use. However the default value can be changed to
3396 permit custom trust settings: one example of this would be to only trust
3397 certificates from a specific "secure" set of CAs.
3399 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
3400 which should be used for version portability: especially since the
3401 verify structure is likely to change more often now.
3403 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
3404 to set them. If not set then assume SSL clients will verify SSL servers
3407 Two new options to the verify program: -untrusted allows a set of
3408 untrusted certificates to be passed in and -purpose which sets the
3409 intended purpose of the certificate. If a purpose is set then the
3410 new chain verify code is used to check extension consistency.
3413 *) Support for the authority information access extension.
3416 *) Modify RSA and DSA PEM read routines to transparently handle
3417 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
3418 public keys in a format compatible with certificate
3419 SubjectPublicKeyInfo structures. Unfortunately there were already
3420 functions called *_PublicKey_* which used various odd formats so
3421 these are retained for compatibility: however the DSA variants were
3422 never in a public release so they have been deleted. Changed dsa/rsa
3423 utilities to handle the new format: note no releases ever handled public
3424 keys so we should be OK.
3426 The primary motivation for this change is to avoid the same fiasco
3427 that dogs private keys: there are several incompatible private key
3428 formats some of which are standard and some OpenSSL specific and
3429 require various evil hacks to allow partial transparent handling and
3430 even then it doesn't work with DER formats. Given the option anything
3431 other than PKCS#8 should be dumped: but the other formats have to
3432 stay in the name of compatibility.
3434 With public keys and the benefit of hindsight one standard format
3435 is used which works with EVP_PKEY, RSA or DSA structures: though
3436 it clearly returns an error if you try to read the wrong kind of key.
3438 Added a -pubkey option to the 'x509' utility to output the public key.
3439 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
3440 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
3441 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
3442 that do the same as the EVP_PKEY_assign_*() except they up the
3443 reference count of the added key (they don't "swallow" the
3447 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
3448 CRLs would fail if the file contained no certificates or no CRLs:
3449 added a new function to read in both types and return the number
3450 read: this means that if none are read it will be an error. The
3451 DER versions of the certificate and CRL reader would always fail
3452 because it isn't possible to mix certificates and CRLs in DER format
3453 without choking one or the other routine. Changed this to just read
3454 a certificate: this is the best we can do. Also modified the code
3455 in apps/verify.c to take notice of return codes: it was previously
3456 attempting to read in certificates from NULL pointers and ignoring
3457 any errors: this is one reason why the cert and CRL reader seemed
3458 to work. It doesn't check return codes from the default certificate
3459 routines: these may well fail if the certificates aren't installed.
3462 *) Code to support otherName option in GeneralName.
3465 *) First update to verify code. Change the verify utility
3466 so it warns if it is passed a self signed certificate:
3467 for consistency with the normal behaviour. X509_verify
3468 has been modified to it will now verify a self signed
3469 certificate if *exactly* the same certificate appears
3470 in the store: it was previously impossible to trust a
3471 single self signed certificate. This means that:
3472 openssl verify ss.pem
3473 now gives a warning about a self signed certificate but
3474 openssl verify -CAfile ss.pem ss.pem
3478 *) For servers, store verify_result in SSL_SESSION data structure
3479 (and add it to external session representation).
3480 This is needed when client certificate verifications fails,
3481 but an application-provided verification callback (set by
3482 SSL_CTX_set_cert_verify_callback) allows accepting the session
3483 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
3484 but returns 1): When the session is reused, we have to set
3485 ssl->verify_result to the appropriate error code to avoid
3487 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
3489 *) Fix a bug in the new PKCS#7 code: it didn't consider the
3490 case in PKCS7_dataInit() where the signed PKCS7 structure
3491 didn't contain any existing data because it was being created.
3492 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
3494 *) Add a salt to the key derivation routines in enc.c. This
3495 forms the first 8 bytes of the encrypted file. Also add a
3496 -S option to allow a salt to be input on the command line.
3499 *) New function X509_cmp(). Oddly enough there wasn't a function
3500 to compare two certificates. We do this by working out the SHA1
3501 hash and comparing that. X509_cmp() will be needed by the trust
3505 *) SSL_get1_session() is like SSL_get_session(), but increments
3506 the reference count in the SSL_SESSION returned.
3507 [Geoff Thorpe <geoff@eu.c2.net>]
3509 *) Fix for 'req': it was adding a null to request attributes.
3510 Also change the X509_LOOKUP and X509_INFO code to handle
3511 certificate auxiliary information.
3514 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
3518 *) Add the possibility to add extra information to the memory leak
3519 detecting output, to form tracebacks, showing from where each
3520 allocation was originated: CRYPTO_push_info("constant string") adds
3521 the string plus current file name and line number to a per-thread
3522 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
3523 is like calling CYRPTO_pop_info() until the stack is empty.
3524 Also updated memory leak detection code to be multi-thread-safe.
3527 *) Add options -text and -noout to pkcs7 utility and delete the
3528 encryption options which never did anything. Update docs.
3531 *) Add options to some of the utilities to allow the pass phrase
3532 to be included on either the command line (not recommended on
3533 OSes like Unix) or read from the environment. Update the
3534 manpages and fix a few bugs.
3537 *) Add a few manpages for some of the openssl commands.
3540 *) Fix the -revoke option in ca. It was freeing up memory twice,
3541 leaking and not finding already revoked certificates.
3544 *) Extensive changes to support certificate auxiliary information.
3545 This involves the use of X509_CERT_AUX structure and X509_AUX
3546 functions. An X509_AUX function such as PEM_read_X509_AUX()
3547 can still read in a certificate file in the usual way but it
3548 will also read in any additional "auxiliary information". By
3549 doing things this way a fair degree of compatibility can be
3550 retained: existing certificates can have this information added
3551 using the new 'x509' options.
3553 Current auxiliary information includes an "alias" and some trust
3554 settings. The trust settings will ultimately be used in enhanced
3555 certificate chain verification routines: currently a certificate
3556 can only be trusted if it is self signed and then it is trusted
3560 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
3561 The problem was that one of the replacement routines had not been working
3562 since SSLeay releases. For now the offending routine has been replaced
3563 with non-optimised assembler. Even so, this now gives around 95%
3564 performance improvement for 1024 bit RSA signs.
3567 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
3568 handling. Most clients have the effective key size in bits equal to
3569 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
3570 A few however don't do this and instead use the size of the decrypted key
3571 to determine the RC2 key length and the AlgorithmIdentifier to determine
3572 the effective key length. In this case the effective key length can still
3573 be 40 bits but the key length can be 168 bits for example. This is fixed
3574 by manually forcing an RC2 key into the EVP_PKEY structure because the
3575 EVP code can't currently handle unusual RC2 key sizes: it always assumes
3576 the key length and effective key length are equal.
3579 *) Add a bunch of functions that should simplify the creation of
3580 X509_NAME structures. Now you should be able to do:
3581 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
3582 and have it automatically work out the correct field type and fill in
3583 the structures. The more adventurous can try:
3584 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
3585 and it will (hopefully) work out the correct multibyte encoding.
3588 *) Change the 'req' utility to use the new field handling and multibyte
3589 copy routines. Before the DN field creation was handled in an ad hoc
3590 way in req, ca, and x509 which was rather broken and didn't support
3591 BMPStrings or UTF8Strings. Since some software doesn't implement
3592 BMPStrings or UTF8Strings yet, they can be enabled using the config file
3593 using the dirstring_type option. See the new comment in the default
3594 openssl.cnf for more info.
3597 *) Make crypto/rand/md_rand.c more robust:
3598 - Assure unique random numbers after fork().
3599 - Make sure that concurrent threads access the global counter and
3600 md serializably so that we never lose entropy in them
3601 or use exactly the same state in multiple threads.
3602 Access to the large state is not always serializable because
3603 the additional locking could be a performance killer, and
3604 md should be large enough anyway.
3607 *) New file apps/app_rand.c with commonly needed functionality
3608 for handling the random seed file.
3610 Use the random seed file in some applications that previously did not:
3612 dsaparam -genkey (which also ignored its '-rand' option),
3615 x509 (when signing).
3616 Except on systems with /dev/urandom, it is crucial to have a random
3617 seed file at least for key creation, DSA signing, and for DH exchanges;
3618 for RSA signatures we could do without one.
3620 gendh and gendsa (unlike genrsa) used to read only the first byte
3621 of each file listed in the '-rand' option. The function as previously
3622 found in genrsa is now in app_rand.c and is used by all programs
3623 that support '-rand'.
3626 *) In RAND_write_file, use mode 0600 for creating files;
3627 don't just chmod when it may be too late.
3630 *) Report an error from X509_STORE_load_locations
3631 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
3634 *) New function ASN1_mbstring_copy() this copies a string in either
3635 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
3636 into an ASN1_STRING type. A mask of permissible types is passed
3637 and it chooses the "minimal" type to use or an error if not type
3641 *) Add function equivalents to the various macros in asn1.h. The old
3642 macros are retained with an M_ prefix. Code inside the library can
3643 use the M_ macros. External code (including the openssl utility)
3644 should *NOT* in order to be "shared library friendly".
3647 *) Add various functions that can check a certificate's extensions
3648 to see if it usable for various purposes such as SSL client,
3649 server or S/MIME and CAs of these types. This is currently
3650 VERY EXPERIMENTAL but will ultimately be used for certificate chain
3651 verification. Also added a -purpose flag to x509 utility to
3652 print out all the purposes.
3655 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
3659 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
3660 for, obtain and decode and extension and obtain its critical flag.
3661 This allows all the necessary extension code to be handled in a
3662 single function call.
3665 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
3666 platforms. See crypto/rc4/rc4_enc.c for further details.
3669 *) New -noout option to asn1parse. This causes no output to be produced
3670 its main use is when combined with -strparse and -out to extract data
3671 from a file (which may not be in ASN.1 format).
3674 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
3675 when producing the local key id.
3676 [Richard Levitte <levitte@stacken.kth.se>]
3678 *) New option -dhparam in s_server. This allows a DH parameter file to be
3679 stated explicitly. If it is not stated then it tries the first server
3680 certificate file. The previous behaviour hard coded the filename
3684 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
3685 a public key to be input or output. For example:
3686 openssl rsa -in key.pem -pubout -out pubkey.pem
3687 Also added necessary DSA public key functions to handle this.
3690 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
3691 in the message. This was handled by allowing
3692 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
3693 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
3695 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
3696 to the end of the strings whereas this didn't. This would cause problems
3697 if strings read with d2i_ASN1_bytes() were later modified.
3698 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
3700 *) Fix for base64 decode bug. When a base64 bio reads only one line of
3701 data and it contains EOF it will end up returning an error. This is
3702 caused by input 46 bytes long. The cause is due to the way base64
3703 BIOs find the start of base64 encoded data. They do this by trying a
3704 trial decode on each line until they find one that works. When they
3705 do a flag is set and it starts again knowing it can pass all the
3706 data directly through the decoder. Unfortunately it doesn't reset
3707 the context it uses. This means that if EOF is reached an attempt
3708 is made to pass two EOFs through the context and this causes the
3709 resulting error. This can also cause other problems as well. As is
3710 usual with these problems it takes *ages* to find and the fix is
3711 trivial: move one line.
3712 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
3714 *) Ugly workaround to get s_client and s_server working under Windows. The
3715 old code wouldn't work because it needed to select() on sockets and the
3716 tty (for keypresses and to see if data could be written). Win32 only
3717 supports select() on sockets so we select() with a 1s timeout on the
3718 sockets and then see if any characters are waiting to be read, if none
3719 are present then we retry, we also assume we can always write data to
3720 the tty. This isn't nice because the code then blocks until we've
3721 received a complete line of data and it is effectively polling the
3722 keyboard at 1s intervals: however it's quite a bit better than not
3723 working at all :-) A dedicated Windows application might handle this
3724 with an event loop for example.
3727 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
3728 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
3729 will be called when RSA_sign() and RSA_verify() are used. This is useful
3730 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
3731 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
3732 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
3733 This necessitated the support of an extra signature type NID_md5_sha1
3734 for SSL signatures and modifications to the SSL library to use it instead
3735 of calling RSA_public_decrypt() and RSA_private_encrypt().
3738 *) Add new -verify -CAfile and -CApath options to the crl program, these
3739 will lookup a CRL issuers certificate and verify the signature in a
3740 similar way to the verify program. Tidy up the crl program so it
3741 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
3742 less strict. It will now permit CRL extensions even if it is not
3743 a V2 CRL: this will allow it to tolerate some broken CRLs.
3746 *) Initialize all non-automatic variables each time one of the openssl
3747 sub-programs is started (this is necessary as they may be started
3748 multiple times from the "OpenSSL>" prompt).
3749 [Lennart Bang, Bodo Moeller]
3751 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
3752 removing all other RSA functionality (this is what NO_RSA does). This
3753 is so (for example) those in the US can disable those operations covered
3754 by the RSA patent while allowing storage and parsing of RSA keys and RSA
3758 *) Non-copying interface to BIO pairs.
3759 (still largely untested)
3762 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
3763 ASCII string. This was handled independently in various places before.
3766 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
3767 UTF8 strings a character at a time.
3770 *) Use client_version from client hello to select the protocol
3771 (s23_srvr.c) and for RSA client key exchange verification
3772 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
3775 *) Add various utility functions to handle SPKACs, these were previously
3776 handled by poking round in the structure internals. Added new function
3777 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
3778 print, verify and generate SPKACs. Based on an original idea from
3779 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
3782 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
3785 *) Allow the config file extension section to be overwritten on the
3786 command line. Based on an original idea from Massimiliano Pala
3787 <madwolf@comune.modena.it>. The new option is called -extensions
3788 and can be applied to ca, req and x509. Also -reqexts to override
3789 the request extensions in req and -crlexts to override the crl extensions
3793 *) Add new feature to the SPKAC handling in ca. Now you can include
3794 the same field multiple times by preceding it by "XXXX." for example:
3797 this is the same syntax as used in the req config file.
3800 *) Allow certificate extensions to be added to certificate requests. These
3801 are specified in a 'req_extensions' option of the req section of the
3802 config file. They can be printed out with the -text option to req but
3803 are otherwise ignored at present.
3806 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
3807 data read consists of only the final block it would not decrypted because
3808 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
3809 A misplaced 'break' also meant the decrypted final block might not be
3810 copied until the next read.
3813 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
3814 a few extra parameters to the DH structure: these will be useful if
3815 for example we want the value of 'q' or implement X9.42 DH.
3818 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
3819 provides hooks that allow the default DSA functions or functions on a
3820 "per key" basis to be replaced. This allows hardware acceleration and
3821 hardware key storage to be handled without major modification to the
3822 library. Also added low level modexp hooks and CRYPTO_EX structure and
3823 associated functions.
3826 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
3827 as "read only": it can't be written to and the buffer it points to will
3828 not be freed. Reading from a read only BIO is much more efficient than
3829 a normal memory BIO. This was added because there are several times when
3830 an area of memory needs to be read from a BIO. The previous method was
3831 to create a memory BIO and write the data to it, this results in two
3832 copies of the data and an O(n^2) reading algorithm. There is a new
3833 function BIO_new_mem_buf() which creates a read only memory BIO from
3834 an area of memory. Also modified the PKCS#7 routines to use read only
3838 *) Bugfix: ssl23_get_client_hello did not work properly when called in
3839 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
3840 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
3841 but a retry condition occured while trying to read the rest.
3844 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
3845 NID_pkcs7_encrypted by default: this was wrong since this should almost
3846 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
3847 the encrypted data type: this is a more sensible place to put it and it
3848 allows the PKCS#12 code to be tidied up that duplicated this
3852 *) Changed obj_dat.pl script so it takes its input and output files on
3853 the command line. This should avoid shell escape redirection problems
3857 *) Initial support for certificate extension requests, these are included
3858 in things like Xenroll certificate requests. Included functions to allow
3859 extensions to be obtained and added.
3862 *) -crlf option to s_client and s_server for sending newlines as
3863 CRLF (as required by many protocols).
3866 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
3868 *) Install libRSAglue.a when OpenSSL is built with RSAref.
3869 [Ralf S. Engelschall]
3871 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
3872 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
3874 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
3878 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
3879 DH parameters/keys (q is lost during that conversion, but the resulting
3880 DH parameters contain its length).
3882 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
3883 much faster than DH_generate_parameters (which creates parameters
3884 where p = 2*q + 1), and also the smaller q makes DH computations
3885 much more efficient (160-bit exponentiation instead of 1024-bit
3886 exponentiation); so this provides a convenient way to support DHE
3887 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
3888 utter importance to use
3889 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
3891 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
3892 when such DH parameters are used, because otherwise small subgroup
3893 attacks may become possible!
3896 *) Avoid memory leak in i2d_DHparams.
3899 *) Allow the -k option to be used more than once in the enc program:
3900 this allows the same encrypted message to be read by multiple recipients.
3903 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
3904 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
3905 it will always use the numerical form of the OID, even if it has a short
3909 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
3910 method only got called if p,q,dmp1,dmq1,iqmp components were present,
3911 otherwise bn_mod_exp was called. In the case of hardware keys for example
3912 no private key components need be present and it might store extra data
3913 in the RSA structure, which cannot be accessed from bn_mod_exp.
3914 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
3915 private key operations.
3918 *) Added support for SPARC Linux.
3921 *) pem_password_cb function type incompatibly changed from
3922 typedef int pem_password_cb(char *buf, int size, int rwflag);
3924 ....(char *buf, int size, int rwflag, void *userdata);
3925 so that applications can pass data to their callbacks:
3926 The PEM[_ASN1]_{read,write}... functions and macros now take an
3927 additional void * argument, which is just handed through whenever
3928 the password callback is called.
3929 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
3931 New function SSL_CTX_set_default_passwd_cb_userdata.
3933 Compatibility note: As many C implementations push function arguments
3934 onto the stack in reverse order, the new library version is likely to
3935 interoperate with programs that have been compiled with the old
3936 pem_password_cb definition (PEM_whatever takes some data that
3937 happens to be on the stack as its last argument, and the callback
3938 just ignores this garbage); but there is no guarantee whatsoever that
3941 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
3942 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
3943 problems not only on Windows, but also on some Unix platforms.
3944 To avoid problematic command lines, these definitions are now in an
3945 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
3946 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
3949 *) MIPS III/IV assembler module is reimplemented.
3952 *) More DES library cleanups: remove references to srand/rand and
3953 delete an unused file.
3956 *) Add support for the the free Netwide assembler (NASM) under Win32,
3957 since not many people have MASM (ml) and it can be hard to obtain.
3958 This is currently experimental but it seems to work OK and pass all
3959 the tests. Check out INSTALL.W32 for info.
3962 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
3963 without temporary keys kept an extra copy of the server key,
3964 and connections with temporary keys did not free everything in case
3968 *) New function RSA_check_key and new openssl rsa option -check
3969 for verifying the consistency of RSA keys.
3970 [Ulf Moeller, Bodo Moeller]
3972 *) Various changes to make Win32 compile work:
3973 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
3974 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
3975 comparison" warnings.
3976 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
3979 *) Add a debugging option to PKCS#5 v2 key generation function: when
3980 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
3981 derived keys are printed to stderr.
3984 *) Copy the flags in ASN1_STRING_dup().
3985 [Roman E. Pavlov <pre@mo.msk.ru>]
3987 *) The x509 application mishandled signing requests containing DSA
3988 keys when the signing key was also DSA and the parameters didn't match.
3990 It was supposed to omit the parameters when they matched the signing key:
3991 the verifying software was then supposed to automatically use the CA's
3992 parameters if they were absent from the end user certificate.
3994 Omitting parameters is no longer recommended. The test was also
3995 the wrong way round! This was probably due to unusual behaviour in
3996 EVP_cmp_parameters() which returns 1 if the parameters match.
3997 This meant that parameters were omitted when they *didn't* match and
3998 the certificate was useless. Certificates signed with 'ca' didn't have
4000 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
4002 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
4003 The interface is as follows:
4004 Applications can use
4005 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
4006 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
4007 "off" is now the default.
4008 The library internally uses
4009 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
4010 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
4011 to disable memory-checking temporarily.
4013 Some inconsistent states that previously were possible (and were
4014 even the default) are now avoided.
4016 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
4017 with each memory chunk allocated; this is occasionally more helpful
4018 than just having a counter.
4020 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
4022 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
4026 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
4027 which largely parallels "options", but is for changing API behaviour,
4028 whereas "options" are about protocol behaviour.
4029 Initial "mode" flags are:
4031 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
4032 a single record has been written.
4033 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
4034 retries use the same buffer location.
4035 (But all of the contents must be
4039 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
4042 *) Fix problems with no-hmac etc.
4043 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
4045 *) New functions RSA_get_default_method(), RSA_set_method() and
4046 RSA_get_method(). These allows replacement of RSA_METHODs without having
4047 to mess around with the internals of an RSA structure.
4050 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
4051 Also really enable memory leak checks in openssl.c and in some
4053 [Chad C. Mulligan, Bodo Moeller]
4055 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
4056 up the length of negative integers. This has now been simplified to just
4057 store the length when it is first determined and use it later, rather
4058 than trying to keep track of where data is copied and updating it to
4060 [Steve Henson, reported by Brien Wheeler
4061 <bwheeler@authentica-security.com>]
4063 *) Add a new function PKCS7_signatureVerify. This allows the verification
4064 of a PKCS#7 signature but with the signing certificate passed to the
4065 function itself. This contrasts with PKCS7_dataVerify which assumes the
4066 certificate is present in the PKCS#7 structure. This isn't always the
4067 case: certificates can be omitted from a PKCS#7 structure and be
4068 distributed by "out of band" means (such as a certificate database).
4071 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
4072 function prototypes in pem.h, also change util/mkdef.pl to add the
4073 necessary function names.
4076 *) mk1mf.pl (used by Windows builds) did not properly read the
4077 options set by Configure in the top level Makefile, and Configure
4078 was not even able to write more than one option correctly.
4079 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
4082 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
4083 file to be loaded from a BIO or FILE pointer. The BIO version will
4084 for example allow memory BIOs to contain config info.
4087 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
4088 Whoever hopes to achieve shared-library compatibility across versions
4089 must use this, not the compile-time macro.
4090 (Exercise 0.9.4: Which is the minimum library version required by
4092 Note: All this applies only to multi-threaded programs, others don't
4096 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
4097 through a BIO pair triggered the default case, i.e.
4098 SSLerr(...,SSL_R_UNKNOWN_STATE).
4101 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
4102 can use the SSL library even if none of the specific BIOs is
4106 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
4107 for the encoded length.
4108 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
4110 *) Add initial documentation of the X509V3 functions.
4113 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
4114 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
4115 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
4116 secure PKCS#8 private key format with a high iteration count.
4119 *) Fix determination of Perl interpreter: A perl or perl5
4120 _directory_ in $PATH was also accepted as the interpreter.
4121 [Ralf S. Engelschall]
4123 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
4124 wrong with it but it was very old and did things like calling
4125 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
4129 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
4130 to use the new extension code.
4133 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
4134 with macros. This should make it easier to change their form, add extra
4135 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
4139 *) Add to configuration table a new entry that can specify an alternative
4140 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
4141 according to Mark Crispin <MRC@Panda.COM>.
4145 *) DES CBC did not update the IV. Weird.
4148 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
4149 Changing the behaviour of the former might break existing programs --
4150 where IV updating is needed, des_ncbc_encrypt can be used.
4153 *) When bntest is run from "make test" it drives bc to check its
4154 calculations, as well as internally checking them. If an internal check
4155 fails, it needs to cause bc to give a non-zero result or make test carries
4156 on without noticing the failure. Fixed.
4159 *) DES library cleanups.
4162 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
4163 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
4164 ciphers. NOTE: although the key derivation function has been verified
4165 against some published test vectors it has not been extensively tested
4166 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
4170 *) Instead of "mkdir -p", which is not fully portable, use new
4171 Perl script "util/mkdir-p.pl".
4174 *) Rewrite the way password based encryption (PBE) is handled. It used to
4175 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
4176 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
4177 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
4178 the 'parameter' field of the AlgorithmIdentifier is passed to the
4179 underlying key generation function so it must do its own ASN1 parsing.
4180 This has also changed the EVP_PBE_CipherInit() function which now has a
4181 'parameter' argument instead of literal salt and iteration count values
4182 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
4185 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
4186 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
4187 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
4188 KEY" because this clashed with PKCS#8 unencrypted string. Since this
4189 value was just used as a "magic string" and not used directly its
4190 value doesn't matter.
4193 *) Introduce some semblance of const correctness to BN. Shame C doesn't
4197 *) "linux-sparc64" configuration (ultrapenguin).
4198 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
4199 "linux-sparc" configuration.
4200 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
4202 *) config now generates no-xxx options for missing ciphers.
4205 *) Support the EBCDIC character set (work in progress).
4206 File ebcdic.c not yet included because it has a different license.
4207 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
4209 *) Support BS2000/OSD-POSIX.
4210 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
4212 *) Make callbacks for key generation use void * instead of char *.
4215 *) Make S/MIME samples compile (not yet tested).
4218 *) Additional typesafe stacks.
4221 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
4225 Changes between 0.9.3 and 0.9.3a [29 May 1999]
4227 *) New configuration variant "sco5-gcc".
4229 *) Updated some demos.
4230 [Sean O Riordain, Wade Scholine]
4232 *) Add missing BIO_free at exit of pkcs12 application.
4235 *) Fix memory leak in conf.c.
4238 *) Updates for Win32 to assembler version of MD5.
4241 *) Set #! path to perl in apps/der_chop to where we found it
4242 instead of using a fixed path.
4245 *) SHA library changes for irix64-mips4-cc.
4248 *) Improvements for VMS support.
4252 Changes between 0.9.2b and 0.9.3 [24 May 1999]
4254 *) Bignum library bug fix. IRIX 6 passes "make test" now!
4255 This also avoids the problems with SC4.2 and unpatched SC5.
4256 [Andy Polyakov <appro@fy.chalmers.se>]
4258 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
4259 These are required because of the typesafe stack would otherwise break
4260 existing code. If old code used a structure member which used to be STACK
4261 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
4262 sk_num or sk_value it would produce an error because the num, data members
4263 are not present in STACK_OF. Now it just produces a warning. sk_set
4264 replaces the old method of assigning a value to sk_value
4265 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
4266 that does this will no longer work (and should use sk_set instead) but
4267 this could be regarded as a "questionable" behaviour anyway.
4270 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
4271 correctly handle encrypted S/MIME data.
4274 *) Change type of various DES function arguments from des_cblock
4275 (which means, in function argument declarations, pointer to char)
4276 to des_cblock * (meaning pointer to array with 8 char elements),
4277 which allows the compiler to do more typechecking; it was like
4278 that back in SSLeay, but with lots of ugly casts.
4280 Introduce new type const_des_cblock.
4283 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
4284 problems: find RecipientInfo structure that matches recipient certificate
4285 and initialise the ASN1 structures properly based on passed cipher.
4288 *) Belatedly make the BN tests actually check the results.
4291 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
4292 to and from BNs: it was completely broken. New compilation option
4293 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
4294 key elements as negative integers.
projects / openssl.git / blob