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 +) Test for certificates which contain unsupported critical extensions.
16 If such a certificate is found during a verify operation it is
17 rejected by default: this behaviour can be overridden by either
18 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
19 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
20 X509_supported_extension() has also been added which returns 1 if a
21 particular extension is supported.
24 +) New functions/macros
26 SSL_CTX_set_msg_callback(ctx, cb)
27 SSL_CTX_set_msg_callback_arg(ctx, arg)
28 SSL_set_msg_callback(ssl, cb)
29 SSL_set_msg_callback_arg(ssl, arg)
31 to request calling a callback function
33 void cb(int write_p, int version, int content_type,
34 const void *buf, size_t len, SSL *ssl, void *arg)
36 whenever a protocol message has been completely received
37 (write_p == 0) or sent (write_p == 1). Here 'version' is the
38 protocol version according to which the SSL library interprets
39 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
40 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
41 the content type as defined in the SSL 3.0/TLS 1.0 protocol
42 specification (change_cipher_spec(20), alert(21), handshake(22)).
43 'buf' and 'len' point to the actual message, 'ssl' to the
44 SSL object, and 'arg' is the application-defined value set by
45 SSL[_CTX]_set_msg_callback_arg().
47 'openssl s_client' and 'openssl s_server' have new '-msg' options
48 to enable a callback that displays all protocol messages.
50 TODO: SSL 2.0, doc/ssl/, doc/apps/
53 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
54 (sent using the client's version number) if client_version is
55 smaller than the protocol version in use. Also change
56 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
57 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
58 the client will at least see that alert.
61 +) Modify the behaviour of EVP cipher functions in similar way to digests
62 to retain compatibility with existing code.
65 +) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
66 compatibility with existing code. In particular the 'ctx' parameter does
67 not have to be to be initialized before the call to EVP_DigestInit() and
68 it is tidied up after a call to EVP_DigestFinal(). New function
69 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
70 EVP_MD_CTX_copy() changed to not require the destination to be
71 initialized valid and new function EVP_MD_CTX_copy_ex() added which
72 requires the destination to be valid.
74 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
75 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
78 +) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
79 so that complete 'Handshake' protocol structures are kept in memory
80 instead of overwriting 'msg_type' and 'length' with 'body' data.
83 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
87 +) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
88 [Massimo Santin via Richard Levitte]
90 +) Major restructuring to the underlying ENGINE code. This includes
91 reduction of linker bloat, separation of pure "ENGINE" manipulation
92 (initialisation, etc) from functionality dealing with implementations
93 of specific crypto iterfaces. This change also introduces integrated
94 support for symmetric ciphers and digest implementations - so ENGINEs
95 can now accelerate these by providing EVP_CIPHER and EVP_MD
96 implementations of their own. This is detailed in crypto/engine/README
97 as it couldn't be adequately described here. However, there are a few
98 API changes worth noting - some RSA, DSA, DH, and RAND functions that
99 were changed in the original introduction of ENGINE code have now
100 reverted back - the hooking from this code to ENGINE is now a good
101 deal more passive and at run-time, operations deal directly with
102 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
103 dereferencing through an ENGINE pointer any more. Also, the ENGINE
104 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
105 they were not being used by the framework as there is no concept of a
106 BIGNUM_METHOD and they could not be generalised to the new
107 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
108 ENGINE_cpy() has been removed as it cannot be consistently defined in
112 +) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
115 +) Change mkdef.pl to sort symbols that get the same entry number,
116 and make sure the automatically generated functions ERR_load_*
117 become part of libeay.num as well.
120 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
121 client receives HelloRequest while in a handshake.
122 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
124 +) New function SSL_renegotiate_pending(). This returns true once
125 renegotiation has been requested (either SSL_renegotiate() call
126 or HelloRequest/ClientHello receveived from the peer) and becomes
127 false once a handshake has been completed.
128 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
129 sends a HelloRequest, but does not ensure that a handshake takes
130 place. SSL_renegotiate_pending() is useful for checking if the
131 client has followed the request.)
134 +) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
135 By default, clients may request session resumption even during
136 renegotiation (if session ID contexts permit); with this option,
137 session resumption is possible only in the first handshake.
140 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
141 should end in 'break', not 'goto end' which circuments various
142 cleanups done in state SSL_ST_OK. But session related stuff
143 must be disabled for SSL_ST_OK in the case that we just sent a
146 Also avoid some overhead by not calling ssl_init_wbio_buffer()
147 before just sending a HelloRequest.
148 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
150 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
151 reveal whether illegal block cipher padding was found or a MAC
152 verification error occured. (Neither SSLerr() codes nor alerts
153 are directly visible to potential attackers, but the information
154 may leak via logfiles.)
156 Similar changes are not required for the SSL 2.0 implementation
157 because the number of padding bytes is sent in clear for SSL 2.0,
158 and the extra bytes are just ignored. However ssl/s2_pkt.c
159 failed to verify that the purported number of padding bytes is in
163 +) Add some demos for certificate and certificate request creation.
166 +) Make maximum certificate chain size accepted from the peer application
167 settable (SSL*_get/set_max_cert_list()), as proposed by
168 "Douglas E. Engert" <deengert@anl.gov>.
171 +) Add support for shared libraries for Unixware-7 and support including
172 shared libraries for OpenUNIX-8 (Boyd Lynn Gerber <gerberb@zenez.com>).
175 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
176 'wristwatch attack' using huge encoding parameters (cf.
177 James H. Manger's CRYPTO 2001 paper). Note that the
178 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
179 encoding parameters and hence was not vulnerable.
182 +) Add a "destroy" handler to ENGINEs that allows structural cleanup to
183 be done prior to destruction. Use this to unload error strings from
184 ENGINEs that load their own error strings. NB: This adds two new API
185 functions to "get" and "set" this destroy handler in an ENGINE.
188 +) Alter all existing ENGINE implementations (except "openssl" and
189 "openbsd") to dynamically instantiate their own error strings. This
190 makes them more flexible to be built both as statically-linked ENGINEs
191 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
192 Also, add stub code to each that makes building them as self-contained
193 shared-libraries easier (see README.ENGINE).
196 +) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
197 implementations into applications that are completely implemented in
198 self-contained shared-libraries. The "dynamic" ENGINE exposes control
199 commands that can be used to configure what shared-library to load and
200 to control aspects of the way it is handled. Also, made an update to
201 the README.ENGINE file that brings its information up-to-date and
202 provides some information and instructions on the "dynamic" ENGINE
203 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
207 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
209 *) Make it possible to unload ranges of ERR strings with a new
210 "ERR_unload_strings" function.
213 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
214 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
215 followed by modular reduction.
216 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
218 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
219 equivalent based on BN_pseudo_rand() instead of BN_rand().
222 +) Add a copy() function to EVP_MD.
225 +) Make EVP_MD routines take a context pointer instead of just the
226 md_data void pointer.
229 +) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
230 that the digest can only process a single chunk of data
231 (typically because it is provided by a piece of
232 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
233 is only going to provide a single chunk of data, and hence the
234 framework needn't accumulate the data for oneshot drivers.
237 +) As with "ERR", make it possible to replace the underlying "ex_data"
238 functions. This change also alters the storage and management of global
239 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
240 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
241 index counters. The API functions that use this state have been changed
242 to take a "class_index" rather than pointers to the class's local STACK
243 and counter, and there is now an API function to dynamically create new
244 classes. This centralisation allows us to (a) plug a lot of the
245 thread-safety problems that existed, and (b) makes it possible to clean
246 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
247 such data would previously have always leaked in application code and
248 workarounds were in place to make the memory debugging turn a blind eye
249 to it. Application code that doesn't use this new function will still
250 leak as before, but their memory debugging output will announce it now
251 rather than letting it slide.
253 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
254 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
255 has a return value to indicate success or failure.
258 +) Make it possible to replace the underlying "ERR" functions such that the
259 global state (2 LHASH tables and 2 locks) is only used by the "default"
260 implementation. This change also adds two functions to "get" and "set"
261 the implementation prior to it being automatically set the first time
262 any other ERR function takes place. Ie. an application can call "get",
263 pass the return value to a module it has just loaded, and that module
264 can call its own "set" function using that value. This means the
265 module's "ERR" operations will use (and modify) the error state in the
266 application and not in its own statically linked copy of OpenSSL code.
269 +) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
270 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
271 the operation, and provides a more encapsulated way for external code
272 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
273 to use these functions rather than manually incrementing the counts.
275 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
278 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
279 This function was broken, as the check for a new client hello message
280 to handle SGC did not allow these large messages.
281 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
284 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
287 +) Add EVP test program.
290 +) Add symmetric cipher support to ENGINE. Expect the API to change!
293 +) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
294 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
295 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
296 These allow a CRL to be built without having to access X509_CRL fields
297 directly. Modify 'ca' application to use new functions.
300 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
301 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
304 *) Rework the configuration and shared library support for Tru64 Unix.
305 The configuration part makes use of modern compiler features and
306 still retains old compiler behavior for those that run older versions
307 of the OS. The shared library support part includes a variant that
308 uses the RPATH feature, and is available through the special
309 configuration target "alpha-cc-rpath", which will never be selected
311 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
313 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
314 with the same message size as in ssl3_get_certificate_request().
315 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
316 messages might inadvertently be reject as too long.
317 [Petr Lampa <lampa@fee.vutbr.cz>]
319 +) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
320 bug workarounds. Rollback attack detection is a security feature.
321 The problem will only arise on OpenSSL servers when TLSv1 is not
322 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
323 Software authors not wanting to support TLSv1 will have special reasons
324 for their choice and can explicitly enable this option.
325 [Bodo Moeller, Lutz Jaenicke]
327 +) Rationalise EVP so it can be extended: don't include a union of
328 cipher/digest structures, add init/cleanup functions. This also reduces
329 the number of header dependencies.
332 +) Make DES key schedule conform to the usual scheme, as well as
333 correcting its structure. This means that calls to DES functions
334 now have to pass a pointer to a des_key_schedule instead of a
335 plain des_key_schedule (which was actually always a pointer
339 +) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
342 *) Modified SSL library such that the verify_callback that has been set
343 specificly for an SSL object with SSL_set_verify() is actually being
344 used. Before the change, a verify_callback set with this function was
345 ignored and the verify_callback() set in the SSL_CTX at the time of
346 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
347 to allow the necessary settings.
350 +) Initial reduction of linker bloat: the use of some functions, such as
351 PEM causes large amounts of unused functions to be linked in due to
352 poor organisation. For example pem_all.c contains every PEM function
353 which has a knock on effect of linking in large amounts of (unused)
354 ASN1 code. Grouping together similar functions and splitting unrelated
355 functions prevents this.
358 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
359 explicitly to NULL, as at least on Solaris 8 this seems not always to be
360 done automatically (in contradiction to the requirements of the C
361 standard). This made problems when used from OpenSSH.
364 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
365 dh->length and always used
367 BN_rand_range(priv_key, dh->p).
369 BN_rand_range() is not necessary for Diffie-Hellman, and this
370 specific range makes Diffie-Hellman unnecessarily inefficient if
371 dh->length (recommended exponent length) is much smaller than the
372 length of dh->p. We could use BN_rand_range() if the order of
373 the subgroup was stored in the DH structure, but we only have
378 BN_rand(priv_key, l, ...)
380 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
386 RSA_eay_public_encrypt
387 RSA_eay_private_decrypt
388 RSA_eay_private_encrypt (signing)
389 RSA_eay_public_decrypt (signature verification)
391 (default implementations for RSA_public_encrypt,
392 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
393 always reject numbers >= n.
396 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
397 to synchronize access to 'locking_thread'. This is necessary on
398 systems where access to 'locking_thread' (an 'unsigned long'
399 variable) is not atomic.
402 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
403 *before* setting the 'crypto_lock_rand' flag. The previous code had
404 a race condition if 0 is a valid thread ID.
405 [Travis Vitek <vitek@roguewave.com>]
407 +) Cleanup of EVP macros.
410 +) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
414 +) Add initial OCSP responder support to ocsp application. The
415 revocation information is handled using the text based index
416 use by the ca application. The responder can either handle
417 requests generated internally, supplied in files (for example
418 via a CGI script) or using an internal minimal server.
421 +) Add configuration choices to get zlib compression for TLS.
424 +) Changes to Kerberos SSL for RFC 2712 compliance:
425 1. Implemented real KerberosWrapper, instead of just using
426 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
427 2. Implemented optional authenticator field of KerberosWrapper.
429 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
430 and authenticator structs; see crypto/krb5/.
432 Generalized Kerberos calls to support multiple Kerberos libraries.
433 [Vern Staats <staatsvr@asc.hpc.mil>,
434 Jeffrey Altman <jaltman@columbia.edu>
437 +) Cause 'openssl speed' to use fully hard-coded DSA keys as it
438 already does with RSA. testdsa.h now has 'priv_key/pub_key'
439 values for each of the key sizes rather than having just
440 parameters (and 'speed' generating keys each time).
443 -) OpenSSL 0.9.6b released [9 July 2001]
445 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
446 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
447 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
448 PRNG state recovery was possible based on the output of
449 one PRNG request appropriately sized to gain knowledge on
450 'md' followed by enough consecutive 1-byte PRNG requests
451 to traverse all of 'state'.
453 1. When updating 'md_local' (the current thread's copy of 'md')
454 during PRNG output generation, hash all of the previous
455 'md_local' value, not just the half used for PRNG output.
457 2. Make the number of bytes from 'state' included into the hash
458 independent from the number of PRNG bytes requested.
460 The first measure alone would be sufficient to avoid
461 Markku-Juhani's attack. (Actually it had never occurred
462 to me that the half of 'md_local' used for chaining was the
463 half from which PRNG output bytes were taken -- I had always
464 assumed that the secret half would be used.) The second
465 measure makes sure that additional data from 'state' is never
466 mixed into 'md_local' in small portions; this heuristically
467 further strengthens the PRNG.
470 +) Speed up EVP routines.
473 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
474 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
475 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
476 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
478 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
479 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
480 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
483 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
485 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
488 *) Fix crypto/bn/asm/mips3.s.
491 *) When only the key is given to "enc", the IV is undefined. Print out
492 an error message in this case.
495 +) Added the OS2-EMX target.
496 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
498 +) Rewrite apps to use NCONF routines instead of the old CONF. New functions
499 to support NCONF routines in extension code. New function CONF_set_nconf()
500 to allow functions which take an NCONF to also handle the old LHASH
501 structure: this means that the old CONF compatible routines can be
502 retained (in particular wrt extensions) without having to duplicate the
503 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
506 *) Handle special case when X509_NAME is empty in X509 printing routines.
509 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
510 positive and less than q.
513 +) Enhance the general user interface with mechanisms for inner control
514 and with possibilities to have yes/no kind of prompts.
517 +) Change all calls to low level digest routines in the library and
518 applications to use EVP. Add missing calls to HMAC_cleanup() and
519 don't assume HMAC_CTX can be copied using memcpy().
520 [Verdon Walker <VWalker@novell.com>, Steve Henson]
522 +) Add the possibility to control engines through control names but with
523 arbitrary arguments instead of just a string.
524 Change the key loaders to take a UI_METHOD instead of a callback
525 function pointer. NOTE: this breaks binary compatibility with earlier
526 versions of OpenSSL [engine].
527 Adapt the nCipher code for these new conditions and add a card insertion
531 +) Enhance the general user interface with mechanisms to better support
532 dialog box interfaces, application-defined prompts, the possibility
533 to use defaults (for example default passwords from somewhere else)
534 and interrupts/cancellations.
537 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
538 used: it isn't thread safe and the add_lock_callback should handle
540 [Paul Rose <Paul.Rose@bridge.com>]
542 *) Verify that incoming data obeys the block size in
543 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
546 +) Tidy up PKCS#12 attribute handling. Add support for the CSP name
547 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
551 [Ulf Möller, Bodo Möller]
553 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
554 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
555 when fixing the server behaviour for backwards-compatible 'client
556 hello' messages. (Note that the attack is impractical against
557 SSL 3.0 and TLS 1.0 anyway because length and version checking
558 means that the probability of guessing a valid ciphertext is
559 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
562 Before 0.9.5, the countermeasure (hide the error by generating a
563 random 'decryption result') did not work properly because
564 ERR_clear_error() was missing, meaning that SSL_get_error() would
565 detect the supposedly ignored error.
567 Both problems are now fixed.
570 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
571 (previously it was 1024).
574 +) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
575 tidy up some unnecessarily weird code in 'sk_new()').
576 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
578 +) Change the key loading routines for ENGINEs to use the same kind
579 callback (pem_password_cb) as all other routines that need this
583 *) Fix for compatibility mode trust settings: ignore trust settings
584 unless some valid trust or reject settings are present.
587 *) Fix for blowfish EVP: its a variable length cipher.
590 +) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
591 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
592 than this minimum value is recommended.
595 +) New random seeder for OpenVMS, using the system process statistics
596 that are easily reachable.
599 +) Windows apparently can't transparently handle global
600 variables defined in DLLs. Initialisations such as:
602 const ASN1_ITEM *it = &ASN1_INTEGER_it;
604 wont compile. This is used by the any applications that need to
605 declare their own ASN1 modules. This was fixed by adding the option
606 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
607 needed for static libraries under Win32.
610 +) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
611 setting of purpose and trust fields. New X509_STORE trust and
612 purpose functions and tidy up setting in other SSL functions.
615 +) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
616 structure. These are inherited by X509_STORE_CTX when it is
617 initialised. This allows various defaults to be set in the
618 X509_STORE structure (such as flags for CRL checking and custom
619 purpose or trust settings) for functions which only use X509_STORE_CTX
620 internally such as S/MIME.
622 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
623 trust settings if they are not set in X509_STORE. This allows X509_STORE
624 purposes and trust (in S/MIME for example) to override any set by default.
626 Add command line options for CRL checking to smime, s_client and s_server
630 +) Initial CRL based revocation checking. If the CRL checking flag(s)
631 are set then the CRL is looked up in the X509_STORE structure and
632 its validity and signature checked, then if the certificate is found
633 in the CRL the verify fails with a revoked error.
635 Various new CRL related callbacks added to X509_STORE_CTX structure.
637 Command line options added to 'verify' application to support this.
639 This needs some additional work, such as being able to handle multiple
640 CRLs with different times, extension based lookup (rather than just
641 by subject name) and ultimately more complete V2 CRL extension
645 +) Add a general user interface API (crypto/ui/). This is designed
646 to replace things like des_read_password and friends (backward
647 compatibility functions using this new API are provided).
648 The purpose is to remove prompting functions from the DES code
649 section as well as provide for prompting through dialog boxes in
650 a window system and the like.
653 *) Fix various bugs related to DSA S/MIME verification. Handle missing
654 parameters in DSA public key structures and return an error in the
655 DSA routines if parameters are absent.
658 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
659 in the current directory if neither $RANDFILE nor $HOME was set.
660 RAND_file_name() in 0.9.6a returned NULL in this case. This has
661 caused some confusion to Windows users who haven't defined $HOME.
662 Thus RAND_file_name() is changed again: e_os.h can define a
663 DEFAULT_HOME, which will be used if $HOME is not set.
664 For Windows, we use "C:"; on other platforms, we still require
665 environment variables.
667 +) Add "ex_data" support to ENGINE so implementations can add state at a
668 per-structure level rather than having to store it globally.
671 +) Make it possible for ENGINE structures to be copied when retrieved by
672 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
673 This causes the "original" ENGINE structure to act like a template,
674 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
675 operational state can be localised to each ENGINE structure, despite the
676 fact they all share the same "methods". New ENGINE structures returned in
677 this case have no functional references and the return value is the single
678 structural reference. This matches the single structural reference returned
679 by ENGINE_by_id() normally, when it is incremented on the pre-existing
683 +) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
684 needs to match any other type at all we need to manually clear the
688 +) Changes to the "openssl engine" utility to include;
689 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
690 about an ENGINE's available control commands.
691 - executing control commands from command line arguments using the
692 '-pre' and '-post' switches. '-post' is only used if '-t' is
693 specified and the ENGINE is successfully initialised. The syntax for
694 the individual commands are colon-separated, for example;
695 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
698 +) New dynamic control command support for ENGINEs. ENGINEs can now
699 declare their own commands (numbers), names (strings), descriptions,
700 and input types for run-time discovery by calling applications. A
701 subset of these commands are implicitly classed as "executable"
702 depending on their input type, and only these can be invoked through
703 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
704 can be based on user input, config files, etc). The distinction is
705 that "executable" commands cannot return anything other than a boolean
706 result and can only support numeric or string input, whereas some
707 discoverable commands may only be for direct use through
708 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
709 pointers, or other custom uses. The "executable" commands are to
710 support parameterisations of ENGINE behaviour that can be
711 unambiguously defined by ENGINEs and used consistently across any
712 OpenSSL-based application. Commands have been added to all the
713 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
714 control over shared-library paths without source code alterations.
717 +) Changed all ENGINE implementations to dynamically allocate their
718 ENGINEs rather than declaring them statically. Apart from this being
719 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
720 this also allows the implementations to compile without using the
721 internal engine_int.h header.
724 +) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
725 'const' value. Any code that should be able to modify a RAND_METHOD
726 should already have non-const pointers to it (ie. they should only
727 modify their own ones).
730 +) Made a variety of little tweaks to the ENGINE code.
731 - "atalla" and "ubsec" string definitions were moved from header files
732 to C code. "nuron" string definitions were placed in variables
733 rather than hard-coded - allowing parameterisation of these values
734 later on via ctrl() commands.
735 - Removed unused "#if 0"'d code.
736 - Fixed engine list iteration code so it uses ENGINE_free() to release
737 structural references.
738 - Constified the RAND_METHOD element of ENGINE structures.
739 - Constified various get/set functions as appropriate and added
740 missing functions (including a catch-all ENGINE_cpy that duplicates
741 all ENGINE values onto a new ENGINE except reference counts/state).
742 - Removed NULL parameter checks in get/set functions. Setting a method
743 or function to NULL is a way of cancelling out a previously set
744 value. Passing a NULL ENGINE parameter is just plain stupid anyway
745 and doesn't justify the extra error symbols and code.
746 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
747 flags from engine_int.h to engine.h.
748 - Changed prototypes for ENGINE handler functions (init(), finish(),
749 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
752 *) Move 'if (!initialized) RAND_poll()' into regions protected by
753 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
754 having multiple threads call RAND_poll() concurrently.
757 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
758 combination of a flag and a thread ID variable.
759 Otherwise while one thread is in ssleay_rand_bytes (which sets the
760 flag), *other* threads can enter ssleay_add_bytes without obeying
761 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
762 that they do not hold after the first thread unsets add_do_not_lock).
765 +) Implement binary inversion algorithm for BN_mod_inverse in addition
766 to the algorithm using long division. The binary algorithm can be
767 used only if the modulus is odd. On 32-bit systems, it is faster
768 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
769 roughly 5-15% for 256-bit moduli), so we use it only for moduli
770 up to 450 bits. In 64-bit environments, the binary algorithm
771 appears to be advantageous for much longer moduli; here we use it
772 for moduli up to 2048 bits.
775 *) Change bctest again: '-x' expressions are not available in all
779 -) OpenSSL 0.9.6a released [5 Apr 2001]
781 *) Fix a couple of memory leaks in PKCS7_dataDecode()
782 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
784 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
785 the default extension for executables, if any. Also, make the perl
786 scripts that use symlink() to test if it really exists and use "cp"
787 if it doesn't. All this made OpenSSL compilable and installable in
791 +) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
792 could not support the combine flag in choice fields.
795 -) Fix for asn1_GetSequence() for indefinite length constructed data.
796 If SEQUENCE is length is indefinite just set c->slen to the total
797 amount of data available.
798 [Steve Henson, reported by shige@FreeBSD.org]
799 [This change does not apply to 0.9.7.]
801 *) Change bctest to avoid here-documents inside command substitution
802 (workaround for FreeBSD /bin/sh bug).
803 For compatibility with Ultrix, avoid shell functions (introduced
804 in the bctest version that searches along $PATH).
807 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
808 with des_encrypt() defined on some operating systems, like Solaris
812 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
813 On the Importance of Eliminating Errors in Cryptographic
814 Computations, J. Cryptology 14 (2001) 2, 101-119,
815 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
818 *) MIPS assembler BIGNUM division bug fix.
821 *) Disabled incorrect Alpha assembler code.
824 -) Fix PKCS#7 decode routines so they correctly update the length
825 after reading an EOC for the EXPLICIT tag.
827 [This change does not apply to 0.9.7.]
829 *) Fix bug in PKCS#12 key generation routines. This was triggered
830 if a 3DES key was generated with a 0 initial byte. Include
831 PKCS12_BROKEN_KEYGEN compilation option to retain the old
832 (but broken) behaviour.
835 *) Enhance bctest to search for a working bc along $PATH and print
837 [Tim Rice <tim@multitalents.net> via Richard Levitte]
839 +) Add a 'copy_extensions' option to the 'ca' utility. This copies
840 extensions from a certificate request to the certificate.
843 +) Allow multiple 'certopt' and 'nameopt' options to be separated
844 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
845 file: this allows the display of the certificate about to be
846 signed to be customised, to allow certain fields to be included
847 or excluded and extension details. The old system didn't display
848 multicharacter strings properly, omitted fields not in the policy
849 and couldn't display additional details such as extensions.
852 *) Fix memory leaks in err.c: free err_data string if necessary;
853 don't write to the wrong index in ERR_set_error_data.
856 +) Function EC_POINTs_mul for simultaneous scalar multiplication
857 of an arbitrary number of elliptic curve points, optionally
858 including the generator defined for the EC_GROUP.
859 EC_POINT_mul is a simple wrapper function for the typical case
860 that the point list has just one item (besides the optional
864 +) First EC_METHODs for curves over GF(p):
866 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
867 operations and provides various method functions that can also
868 operate with faster implementations of modular arithmetic.
870 EC_GFp_mont_method() reuses most functions that are part of
871 EC_GFp_simple_method, but uses Montgomery arithmetic.
873 [Bodo Moeller; point addition and point doubling
874 implementation directly derived from source code provided by
875 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
877 +) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
880 Curves are EC_GROUP objects (with an optional group generator)
881 based on EC_METHODs that are built into the library.
883 Points are EC_POINT objects based on EC_GROUP objects.
885 Most of the framework would be able to handle curves over arbitrary
886 finite fields, but as there are no obvious types for fields other
887 than GF(p), some functions are limited to that for now.
890 +) Add the -HTTP option to s_server. It is similar to -WWW, but requires
891 that the file contains a complete HTTP response.
894 +) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
895 change the def and num file printf format specifier from "%-40sXXX"
896 to "%-39s XXX". The latter will always guarantee a space after the
897 field while the former will cause them to run together if the field
898 is 40 of more characters long.
901 +) Constify the cipher and digest 'method' functions and structures
902 and modify related functions to take constant EVP_MD and EVP_CIPHER
906 *) Implement ssl23_peek (analogous to ssl23_read), which previously
910 *) Replace rdtsc with _emit statements for VC++ version 5.
911 [Jeremy Cooper <jeremy@baymoo.org>]
913 +) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
914 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
917 +) Modify EVP_Digest*() routines so they now return values. Although the
918 internal software routines can never fail additional hardware versions
922 +) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
924 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
925 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
928 ERR_R_NESTED_ASN1_ERROR
930 ERR_R_MISSING_ASN1_EOS
931 were 4 .. 9, conflicting with
932 ERR_LIB_RSA (= ERR_R_RSA_LIB)
934 ERR_LIB_PEM (= ERR_R_PEM_LIB).
935 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
937 Add new error code 'ERR_R_INTERNAL_ERROR'.
940 +) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
944 +) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
945 sets the subject name for a new request or supersedes the
946 subject name in a given request. Formats that can be parsed are
947 'CN=Some Name, OU=myOU, C=IT'
949 'CN=Some Name/OU=myOU/C=IT'.
951 Add options '-batch' and '-verbose' to 'openssl req'.
952 [Massimiliano Pala <madwolf@hackmasters.net>]
954 *) Make it possible to reuse SSLv2 sessions.
957 +) Introduce the possibility to access global variables through
958 functions on platform were that's the best way to handle exporting
959 global variables in shared libraries. To enable this functionality,
960 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
961 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
962 is normally done by Configure or something similar).
964 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
965 in the source file (foo.c) like this:
967 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
968 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
970 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
971 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
973 OPENSSL_DECLARE_GLOBAL(int,foo);
974 #define foo OPENSSL_GLOBAL_REF(foo)
975 OPENSSL_DECLARE_GLOBAL(double,bar);
976 #define bar OPENSSL_GLOBAL_REF(bar)
978 The #defines are very important, and therefore so is including the
979 header file everywhere where the defined globals are used.
981 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
982 of ASN.1 items, but that structure is a bit different.
984 The largest change is in util/mkdef.pl which has been enhanced with
985 better and easier to understand logic to choose which symbols should
986 go into the Windows .def files as well as a number of fixes and code
987 cleanup (among others, algorithm keywords are now sorted
988 lexicographically to avoid constant rewrites).
991 *) In copy_email() check for >= 0 as a return value for
992 X509_NAME_get_index_by_NID() since 0 is a valid index.
993 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
995 +) In BN_div() keep a copy of the sign of 'num' before writing the
996 result to 'rm' because if rm==num the value will be overwritten
997 and produce the wrong result if 'num' is negative: this caused
998 problems with BN_mod() and BN_nnmod().
1001 +) Function OCSP_request_verify(). This checks the signature on an
1002 OCSP request and verifies the signer certificate. The signer
1003 certificate is just checked for a generic purpose and OCSP request
1007 +) Add OCSP_check_validity() function to check the validity of OCSP
1008 responses. OCSP responses are prepared in real time and may only
1009 be a few seconds old. Simply checking that the current time lies
1010 between thisUpdate and nextUpdate max reject otherwise valid responses
1011 caused by either OCSP responder or client clock inaccuracy. Instead
1012 we allow thisUpdate and nextUpdate to fall within a certain period of
1013 the current time. The age of the response can also optionally be
1014 checked. Two new options -validity_period and -status_age added to
1018 +) If signature or public key algorithm is unrecognized print out its
1019 OID rather that just UNKNOWN.
1022 *) Avoid coredump with unsupported or invalid public keys by checking if
1023 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
1024 PKCS7_verify() fails with non detached data.
1027 +) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
1028 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
1029 ID to be generated from the issuer certificate alone which can then be
1030 passed to OCSP_id_issuer_cmp().
1033 +) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
1034 ASN1 modules to export functions returning ASN1_ITEM pointers
1035 instead of the ASN1_ITEM structures themselves. This adds several
1036 new macros which allow the underlying ASN1 function/structure to
1037 be accessed transparently. As a result code should not use ASN1_ITEM
1038 references directly (such as &X509_it) but instead use the relevant
1039 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
1040 use of the new ASN1 code on platforms where exporting structures
1041 is problematical (for example in shared libraries) but exporting
1042 functions returning pointers to structures is not.
1045 +) Add support for overriding the generation of SSL/TLS session IDs.
1046 These callbacks can be registered either in an SSL_CTX or per SSL.
1047 The purpose of this is to allow applications to control, if they wish,
1048 the arbitrary values chosen for use as session IDs, particularly as it
1049 can be useful for session caching in multiple-server environments. A
1050 command-line switch for testing this (and any client code that wishes
1051 to use such a feature) has been added to "s_server".
1052 [Geoff Thorpe, Lutz Jaenicke]
1054 +) Modify mkdef.pl to recognise and parse preprocessor conditionals
1055 of the form '#if defined(...) || defined(...) || ...' and
1056 '#if !defined(...) && !defined(...) && ...'. This also avoids
1057 the growing number of special cases it was previously handling.
1060 *) Don't use getenv in library functions when run as setuid/setgid.
1061 New function OPENSSL_issetugid().
1064 +) Make all configuration macros available for application by making
1065 sure they are available in opensslconf.h, by giving them names starting
1066 with "OPENSSL_" to avoid conflicts with other packages and by making
1067 sure e_os2.h will cover all platform-specific cases together with
1069 Additionally, it is now possible to define configuration/platform-
1070 specific names (called "system identities"). In the C code, these
1071 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
1072 macro with the name beginning with "OPENSSL_SYS_", which is determined
1073 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
1077 +) New option -set_serial to 'req' and 'x509' this allows the serial
1078 number to use to be specified on the command line. Previously self
1079 signed certificates were hard coded with serial number 0 and the
1080 CA options of 'x509' had to use a serial number in a file which was
1084 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
1085 due to incorrect handling of multi-threading:
1087 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
1089 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
1091 3. Count how many times MemCheck_off() has been called so that
1092 nested use can be treated correctly. This also avoids
1093 inband-signalling in the previous code (which relied on the
1094 assumption that thread ID 0 is impossible).
1097 +) New options to 'ca' utility to support V2 CRL entry extensions.
1098 Currently CRL reason, invalidity date and hold instruction are
1099 supported. Add new CRL extensions to V3 code and some new objects.
1102 *) Add "-rand" option also to s_client and s_server.
1105 +) New function EVP_CIPHER_CTX_set_padding() this is used to
1106 disable standard block padding (aka PKCS#5 padding) in the EVP
1107 API, which was previously mandatory. This means that the data is
1108 not padded in any way and so the total length much be a multiple
1109 of the block size, otherwise an error occurs.
1112 +) Initial (incomplete) OCSP SSL support.
1115 *) Fix CPU detection on Irix 6.x.
1116 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
1117 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
1119 +) New function OCSP_parse_url(). This splits up a URL into its host,
1120 port and path components: primarily to parse OCSP URLs. New -url
1121 option to ocsp utility.
1124 +) New nonce behavior. The return value of OCSP_check_nonce() now
1125 reflects the various checks performed. Applications can decide
1126 whether to tolerate certain situations such as an absent nonce
1127 in a response when one was present in a request: the ocsp application
1128 just prints out a warning. New function OCSP_add1_basic_nonce()
1129 this is to allow responders to include a nonce in a response even if
1130 the request is nonce-less.
1133 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
1136 [This change does not apply to 0.9.7.]
1138 *) Use the cached encoding of an X509_NAME structure rather than
1139 copying it. This is apparently the reason for the libsafe "errors"
1140 but the code is actually correct.
1143 +) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
1144 skipped when using openssl x509 multiple times on a single input file,
1145 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
1148 +) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
1149 set string type: to handle setting ASN1_TIME structures. Fix ca
1150 utility to correctly initialize revocation date of CRLs.
1153 +) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
1154 the clients preferred ciphersuites and rather use its own preferences.
1155 Should help to work around M$ SGC (Server Gated Cryptography) bug in
1156 Internet Explorer by ensuring unchanged hash method during stepup.
1157 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
1160 +) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
1161 to aes and add a new 'exist' option to print out symbols that don't
1165 +) Additional options to ocsp utility to allow flags to be set and
1166 additional certificates supplied.
1169 +) Add the option -VAfile to 'openssl ocsp', so the user can give the
1170 OCSP client a number of certificate to only verify the response
1174 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
1175 Bleichenbacher's DSA attack.
1176 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
1177 to be set and top=0 forces the highest bit to be set; top=-1 is new
1178 and leaves the highest bit random.
1179 [Ulf Moeller, Bodo Moeller]
1181 +) Update Rijndael code to version 3.0 and change EVP AES ciphers to
1182 handle the new API. Currently only ECB, CBC modes supported. Add new
1183 AES OIDs. Add TLS AES ciphersuites as described in the "AES Ciphersuites
1184 for TLS" draft-ietf-tls-ciphersuite-03.txt.
1185 [Ben Laurie, Steve Henson]
1187 *) In the NCONF_...-based implementations for CONF_... queries
1188 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
1189 a temporary CONF structure with the data component set to NULL
1190 (which gives segmentation faults in lh_retrieve).
1191 Instead, use NULL for the CONF pointer in CONF_get_string and
1192 CONF_get_number (which may use environment variables) and directly
1193 return NULL from CONF_get_section.
1196 *) Fix potential buffer overrun for EBCDIC.
1199 +) New function OCSP_copy_nonce() to copy nonce value (if present) from
1200 request to response.
1203 +) Functions for OCSP responders. OCSP_request_onereq_count(),
1204 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
1205 extract information from a certificate request. OCSP_response_create()
1206 creates a response and optionally adds a basic response structure.
1207 OCSP_basic_add1_status() adds a complete single response to a basic
1208 response and returns the OCSP_SINGLERESP structure just added (to allow
1209 extensions to be included for example). OCSP_basic_add1_cert() adds a
1210 certificate to a basic response and OCSP_basic_sign() signs a basic
1211 response with various flags. New helper functions ASN1_TIME_check()
1212 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
1213 (converts ASN1_TIME to GeneralizedTime).
1216 +) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
1217 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
1218 structure from a certificate. X509_pubkey_digest() digests the public_key
1219 contents: this is used in various key identifiers.
1222 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
1223 keyUsage if basicConstraints absent for a CA.
1226 *) Make SMIME_write_PKCS7() write mail header values with a format that
1227 is more generally accepted (no spaces before the semicolon), since
1228 some programs can't parse those values properly otherwise. Also make
1229 sure BIO's that break lines after each write do not create invalid
1233 +) Make sk_sort() tolerate a NULL argument.
1234 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
1236 +) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
1237 passed by the function are trusted implicitly. If any of them signed the
1238 response then it is assumed to be valid and is not verified.
1241 -) Make the CRL encoding routines work with empty SEQUENCE OF. The
1242 macros previously used would not encode an empty SEQUENCE OF
1243 and break the signature.
1245 [This change does not apply to 0.9.7.]
1247 *) Zero the premaster secret after deriving the master secret in
1251 +) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
1252 to data. This was previously part of the PKCS7 ASN1 code. This
1253 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
1254 [Steve Henson, reported by Kenneth R. Robinette
1255 <support@securenetterm.com>]
1257 +) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
1258 routines: without these tracing memory leaks is very painful.
1259 Fix leaks in PKCS12 and PKCS7 routines.
1262 *) Add some EVP_add_digest_alias registrations (as found in
1263 OpenSSL_add_all_digests()) to SSL_library_init()
1264 aka OpenSSL_add_ssl_algorithms(). This provides improved
1265 compatibility with peers using X.509 certificates
1266 with unconventional AlgorithmIdentifier OIDs.
1269 *) Fix for Irix with NO_ASM.
1270 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
1272 *) ./config script fixes.
1273 [Ulf Moeller, Richard Levitte]
1275 +) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
1276 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
1277 effectively meant GeneralizedTime would never be used. Now it
1278 is initialised to -1 but X509_time_adj() now has to check the value
1279 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
1280 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
1281 [Steve Henson, reported by Kenneth R. Robinette
1282 <support@securenetterm.com>]
1284 +) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
1285 result in a zero length in the ASN1_INTEGER structure which was
1286 not consistent with the structure when d2i_ASN1_INTEGER() was used
1287 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
1288 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
1289 where it did not print out a minus for negative ASN1_INTEGER.
1292 *) Fix 'openssl passwd -1'.
1295 +) Add summary printout to ocsp utility. The various functions which
1296 convert status values to strings have been renamed to:
1297 OCSP_response_status_str(), OCSP_cert_status_str() and
1298 OCSP_crl_reason_str() and are no longer static. New options
1299 to verify nonce values and to disable verification. OCSP response
1300 printout format cleaned up.
1303 +) Add additional OCSP certificate checks. These are those specified
1304 in RFC2560. This consists of two separate checks: the CA of the
1305 certificate being checked must either be the OCSP signer certificate
1306 or the issuer of the OCSP signer certificate. In the latter case the
1307 OCSP signer certificate must contain the OCSP signing extended key
1308 usage. This check is performed by attempting to match the OCSP
1309 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
1310 in the OCSP_CERTID structures of the response.
1313 +) Initial OCSP certificate verification added to OCSP_basic_verify()
1314 and related routines. This uses the standard OpenSSL certificate
1315 verify routines to perform initial checks (just CA validity) and
1316 to obtain the certificate chain. Then additional checks will be
1317 performed on the chain. Currently the root CA is checked to see
1318 if it is explicitly trusted for OCSP signing. This is used to set
1319 a root CA as a global signing root: that is any certificate that
1320 chains to that CA is an acceptable OCSP signing certificate.
1323 +) New '-extfile ...' option to 'openssl ca' for reading X.509v3
1324 extensions from a separate configuration file.
1325 As when reading extensions from the main configuration file,
1326 the '-extensions ...' option may be used for specifying the
1328 [Massimiliano Pala <madwolf@comune.modena.it>]
1330 *) Change PKCS12_key_gen_asc() so it can cope with non null
1331 terminated strings whose length is passed in the passlen
1332 parameter, for example from PEM callbacks. This was done
1333 by adding an extra length parameter to asc2uni().
1334 [Steve Henson, reported by <oddissey@samsung.co.kr>]
1336 +) New OCSP utility. Allows OCSP requests to be generated or
1337 read. The request can be sent to a responder and the output
1338 parsed, outputed or printed in text form. Not complete yet:
1339 still needs to check the OCSP response validity.
1342 +) New subcommands for 'openssl ca':
1343 'openssl ca -status <serial>' prints the status of the cert with
1344 the given serial number (according to the index file).
1345 'openssl ca -updatedb' updates the expiry status of certificates
1347 [Massimiliano Pala <madwolf@comune.modena.it>]
1349 +) New '-newreq-nodes' command option to CA.pl. This is like
1350 '-newreq', but calls 'openssl req' with the '-nodes' option
1351 so that the resulting key is not encrypted.
1352 [Damien Miller <djm@mindrot.org>]
1354 +) New configuration for the GNU Hurd.
1355 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
1357 +) Initial code to implement OCSP basic response verify. This
1358 is currently incomplete. Currently just finds the signer's
1359 certificate and verifies the signature on the response.
1362 +) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
1363 value of OPENSSLDIR. This is available via the new '-d' option
1364 to 'openssl version', and is also included in 'openssl version -a'.
1367 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
1368 call failed, free the DSA structure.
1371 +) Allowing defining memory allocation callbacks that will be given
1372 file name and line number information in additional arguments
1373 (a const char* and an int). The basic functionality remains, as
1374 well as the original possibility to just replace malloc(),
1375 realloc() and free() by functions that do not know about these
1376 additional arguments. To register and find out the current
1377 settings for extended allocation functions, the following
1378 functions are provided:
1380 CRYPTO_set_mem_ex_functions
1381 CRYPTO_set_locked_mem_ex_functions
1382 CRYPTO_get_mem_ex_functions
1383 CRYPTO_get_locked_mem_ex_functions
1385 These work the same way as CRYPTO_set_mem_functions and friends.
1386 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
1387 extended allocation function is enabled.
1388 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
1389 a conventional allocation function is enabled.
1390 [Richard Levitte, Bodo Moeller]
1392 *) Fix to uni2asc() to cope with zero length Unicode strings.
1393 These are present in some PKCS#12 files.
1396 +) Finish off removing the remaining LHASH function pointer casts.
1397 There should no longer be any prototype-casting required when using
1398 the LHASH abstraction, and any casts that remain are "bugs". See
1399 the callback types and macros at the head of lhash.h for details
1400 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
1403 +) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
1404 If /dev/[u]random devices are not available or do not return enough
1405 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
1407 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
1408 /etc/entropy will be queried once each in this sequence, quering stops
1409 when enough entropy was collected without querying more sockets.
1412 +) Change the Unix RAND_poll() variant to be able to poll several
1413 random devices, as specified by DEVRANDOM, until a sufficient amount
1414 of data has been collected. We spend at most 10 ms on each file
1415 (select timeout) and read in non-blocking mode. DEVRANDOM now
1416 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
1417 (previously it was just the string "/dev/urandom"), so on typical
1418 platforms the 10 ms delay will never occur.
1419 Also separate out the Unix variant to its own file, rand_unix.c.
1420 For VMS, there's a currently-empty rand_vms.c.
1423 +) Move OCSP client related routines to ocsp_cl.c. These
1424 provide utility functions which an application needing
1425 to issue a request to an OCSP responder and analyse the
1426 response will typically need: as opposed to those which an
1427 OCSP responder itself would need which will be added later.
1429 OCSP_request_sign() signs an OCSP request with an API similar
1430 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
1431 response. OCSP_response_get1_basic() extracts basic response
1432 from response. OCSP_resp_find_status(): finds and extracts status
1433 information from an OCSP_CERTID structure (which will be created
1434 when the request structure is built). These are built from lower
1435 level functions which work on OCSP_SINGLERESP structures but
1436 wont normally be used unless the application wishes to examine
1437 extensions in the OCSP response for example.
1439 Replace nonce routines with a pair of functions.
1440 OCSP_request_add1_nonce() adds a nonce value and optionally
1441 generates a random value. OCSP_check_nonce() checks the
1442 validity of the nonce in an OCSP response.
1445 +) Change function OCSP_request_add() to OCSP_request_add0_id().
1446 This doesn't copy the supplied OCSP_CERTID and avoids the
1447 need to free up the newly created id. Change return type
1448 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
1449 This can then be used to add extensions to the request.
1450 Deleted OCSP_request_new(), since most of its functionality
1451 is now in OCSP_REQUEST_new() (and the case insensitive name
1452 clash) apart from the ability to set the request name which
1453 will be added elsewhere.
1456 +) Update OCSP API. Remove obsolete extensions argument from
1457 various functions. Extensions are now handled using the new
1458 OCSP extension code. New simple OCSP HTTP function which
1459 can be used to send requests and parse the response.
1462 +) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
1463 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
1464 uses the special reorder version of SET OF to sort the attributes
1465 and reorder them to match the encoded order. This resolves a long
1466 standing problem: a verify on a PKCS7 structure just after signing
1467 it used to fail because the attribute order did not match the
1468 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
1469 it uses the received order. This is necessary to tolerate some broken
1470 software that does not order SET OF. This is handled by encoding
1471 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
1472 to produce the required SET OF.
1475 +) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
1476 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
1477 files to get correct declarations of the ASN.1 item variables.
1480 +) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
1481 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
1482 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
1483 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
1484 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
1485 ASN1_ITEM and no wrapper functions.
1488 +) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
1489 replace the old function pointer based I/O routines. Change most of
1490 the *_d2i_bio() and *_d2i_fp() functions to use these.
1493 +) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
1494 lines, recognice more "algorithms" that can be deselected, and make
1495 it complain about algorithm deselection that isn't recognised.
1498 +) New ASN1 functions to handle dup, sign, verify, digest, pack and
1499 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
1500 to use new functions. Add NO_ASN1_OLD which can be set to remove
1501 some old style ASN1 functions: this can be used to determine if old
1502 code will still work when these eventually go away.
1505 +) New extension functions for OCSP structures, these follow the
1506 same conventions as certificates and CRLs.
1509 +) New function X509V3_add1_i2d(). This automatically encodes and
1510 adds an extension. Its behaviour can be customised with various
1511 flags to append, replace or delete. Various wrappers added for
1512 certifcates and CRLs.
1515 +) Fix to avoid calling the underlying ASN1 print routine when
1516 an extension cannot be parsed. Correct a typo in the
1517 OCSP_SERVICELOC extension. Tidy up print OCSP format.
1520 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
1521 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
1522 when writing a 32767 byte record.
1523 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
1525 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
1526 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
1528 (RSA objects have a reference count access to which is protected
1529 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
1530 so they are meant to be shared between threads.)
1531 [Bodo Moeller, Geoff Thorpe; original patch submitted by
1532 "Reddie, Steven" <Steven.Reddie@ca.com>]
1534 +) Make mkdef.pl parse some of the ASN1 macros and add apropriate
1535 entries for variables.
1538 *) Fix a deadlock in CRYPTO_mem_leaks().
1541 +) Add functionality to apps/openssl.c for detecting locking
1542 problems: As the program is single-threaded, all we have
1543 to do is register a locking callback using an array for
1544 storing which locks are currently held by the program.
1547 +) Use a lock around the call to CRYPTO_get_ex_new_index() in
1548 SSL_get_ex_data_X509_STORE_idx(), which is used in
1549 ssl_verify_cert_chain() and thus can be called at any time
1550 during TLS/SSL handshakes so that thread-safety is essential.
1551 Unfortunately, the ex_data design is not at all suited
1552 for multi-threaded use, so it probably should be abolished.
1555 +) Added Broadcom "ubsec" ENGINE to OpenSSL.
1556 [Broadcom, tweaked and integrated by Geoff Thorpe]
1558 +) Move common extension printing code to new function
1559 X509V3_print_extensions(). Reorganise OCSP print routines and
1560 implement some needed OCSP ASN1 functions. Add OCSP extensions.
1563 +) New function X509_signature_print() to remove duplication in some
1567 +) Add a special meaning when SET OF and SEQUENCE OF flags are both
1568 set (this was treated exactly the same as SET OF previously). This
1569 is used to reorder the STACK representing the structure to match the
1570 encoding. This will be used to get round a problem where a PKCS7
1571 structure which was signed could not be verified because the STACK
1572 order did not reflect the encoded order.
1575 +) Reimplement the OCSP ASN1 module using the new code.
1578 +) Update the X509V3 code to permit the use of an ASN1_ITEM structure
1579 for its ASN1 operations. The old style function pointers still exist
1580 for now but they will eventually go away.
1583 +) Merge in replacement ASN1 code from the ASN1 branch. This almost
1584 completely replaces the old ASN1 functionality with a table driven
1585 encoder and decoder which interprets an ASN1_ITEM structure describing
1586 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
1587 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
1588 has also been converted to the new form.
1591 +) Change BN_mod_exp_recp so that negative moduli are tolerated
1592 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
1593 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
1594 for negative moduli.
1597 +) Fix BN_uadd and BN_usub: Always return non-negative results instead
1598 of not touching the result's sign bit.
1601 +) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
1605 +) Changed the LHASH code to use prototypes for callbacks, and created
1606 macros to declare and implement thin (optionally static) functions
1607 that provide type-safety and avoid function pointer casting for the
1608 type-specific callbacks.
1611 *) Use better test patterns in bntest.
1614 +) Added Kerberos Cipher Suites to be used with TLS, as written in
1616 [Veers Staats <staatsvr@asc.hpc.mil>,
1617 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
1619 *) rand_win.c fix for Borland C.
1622 *) BN_rshift bugfix for n == 0.
1625 +) Reformat the FAQ so the different questions and answers can be divided
1626 in sections depending on the subject.
1629 +) Have the zlib compression code load ZLIB.DLL dynamically under
1633 +) New function BN_mod_sqrt for computing square roots modulo a prime
1634 (using the probabilistic Tonelli-Shanks algorithm unless
1635 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
1636 be handled deterministically).
1637 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
1639 *) Add a 'bctest' script that checks for some known 'bc' bugs
1640 so that 'make test' does not abort just because 'bc' is broken.
1643 *) Store verify_result within SSL_SESSION also for client side to
1644 avoid potential security hole. (Re-used sessions on the client side
1645 always resulted in verify_result==X509_V_OK, not using the original
1646 result of the server certificate verification.)
1649 +) Make BN_mod_inverse faster by explicitly handling small quotients
1650 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
1651 512 bits], about 30% for larger ones [1024 or 2048 bits].)
1654 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
1655 SSL3_RT_APPLICATION_DATA, return 0.
1656 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
1660 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
1661 releases, have been re-implemented by renaming the previous
1662 implementations of ssl2_read and ssl3_read to ssl2_read_internal
1663 and ssl3_read_internal, respectively, and adding 'peek' parameters
1664 to them. The new ssl[23]_{read,peek} functions are calls to
1665 ssl[23]_read_internal with the 'peek' flag set appropriately.
1666 A 'peek' parameter has also been added to ssl3_read_bytes, which
1667 does the actual work for ssl3_read_internal.
1670 +) New function BN_kronecker.
1673 +) Fix BN_gcd so that it works on negative inputs; the result is
1674 positive unless both parameters are zero.
1675 Previously something reasonably close to an infinite loop was
1676 possible because numbers could be growing instead of shrinking
1677 in the implementation of Euclid's algorithm.
1680 +) Fix BN_is_word() and BN_is_one() macros to take into account the
1681 sign of the number in question.
1683 Fix BN_is_word(a,w) to work correctly for w == 0.
1685 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
1686 because its test if the absolute value of 'a' equals 'w'.
1687 Note that BN_abs_is_word does *not* handle w == 0 reliably;
1688 it exists mostly for use in the implementations of BN_is_zero(),
1689 BN_is_one(), and BN_is_word().
1692 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
1693 the method-specific "init()" handler. Also clean up ex_data after
1694 calling the method-specific "finish()" handler. Previously, this was
1695 happening the other way round.
1698 +) New function BN_swap.
1701 +) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
1702 the exponentiation functions are more likely to produce reasonable
1703 results on negative inputs.
1706 +) Change BN_mod_mul so that the result is always non-negative.
1707 Previously, it could be negative if one of the factors was negative;
1708 I don't think anyone really wanted that behaviour.
1711 +) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
1712 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
1713 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
1714 and add new functions:
1723 BN_mod_lshift1_quick
1727 These functions always generate non-negative results.
1729 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
1730 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
1732 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
1733 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
1734 be reduced modulo m.
1735 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
1737 +) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
1738 was actually never needed) and in BN_mul(). The removal in BN_mul()
1739 required a small change in bn_mul_part_recursive() and the addition
1740 of the functions bn_cmp_part_words(), bn_sub_part_words() and
1741 bn_add_part_words(), which do the same thing as bn_cmp_words(),
1742 bn_sub_words() and bn_add_words() except they take arrays with
1746 +) In 'openssl passwd', verify passwords read from the terminal
1747 unless the '-salt' option is used (which usually means that
1748 verification would just waste user's time since the resulting
1749 hash is going to be compared with some given password hash)
1750 or the new '-noverify' option is used.
1752 This is an incompatible change, but it does not affect
1753 non-interactive use of 'openssl passwd' (passwords on the command
1754 line, '-stdin' option, '-in ...' option) and thus should not
1758 +) Remove all references to RSAref, since there's no more need for it.
1761 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
1762 The previous value, 12, was not always sufficient for BN_mod_exp().
1765 +) Make DSO load along a path given through an environment variable
1766 (SHLIB_PATH) with shl_load().
1769 +) Constify the ENGINE code as a result of BIGNUM constification.
1770 Also constify the RSA code and most things related to it. In a
1771 few places, most notable in the depth of the ASN.1 code, ugly
1772 casts back to non-const were required (to be solved at a later
1776 +) Make it so the openssl application has all engines loaded by default.
1779 +) Constify the BIGNUM routines a little more.
1782 *) Make sure that shared libraries get the internal name engine with
1783 the full version number and not just 0. This should mark the
1784 shared libraries as not backward compatible. Of course, this should
1785 be changed again when we can guarantee backward binary compatibility.
1788 +) Add the following functions:
1790 ENGINE_load_cswift()
1792 ENGINE_load_atalla()
1794 ENGINE_load_builtin_engines()
1796 That way, an application can itself choose if external engines that
1797 are built-in in OpenSSL shall ever be used or not. The benefit is
1798 that applications won't have to be linked with libdl or other dso
1799 libraries unless it's really needed.
1801 Changed 'openssl engine' to load all engines on demand.
1802 Changed the engine header files to avoid the duplication of some
1803 declarations (they differed!).
1806 +) 'openssl engine' can now list capabilities.
1809 +) Better error reporting in 'openssl engine'.
1812 +) Never call load_dh_param(NULL) in s_server.
1815 +) Add engine application. It can currently list engines by name and
1816 identity, and test if they are actually available.
1819 +) Add support for shared libraries under Irix.
1820 [Albert Chin-A-Young <china@thewrittenword.com>]
1822 +) Improve RPM specification file by forcing symbolic linking and making
1823 sure the installed documentation is also owned by root.root.
1824 [Damien Miller <djm@mindrot.org>]
1826 +) Add configuration option to build on Linux on both big-endian and
1828 [Ralf Baechle <ralf@uni-koblenz.de>]
1830 +) Give the OpenSSL applications more possibilities to make use of
1831 keys (public as well as private) handled by engines.
1834 +) Add OCSP code that comes from CertCo.
1837 +) Add VMS support for the Rijndael code.
1840 +) Added untested support for Nuron crypto accelerator.
1843 +) Add support for external cryptographic devices. This code was
1844 previously distributed separately as the "engine" branch.
1845 [Geoff Thorpe, Richard Levitte]
1847 +) Rework the filename-translation in the DSO code. It is now possible to
1848 have far greater control over how a "name" is turned into a filename
1849 depending on the operating environment and any oddities about the
1850 different shared library filenames on each system.
1853 +) Support threads on FreeBSD-elf in Configure.
1856 +) Add the possibility to create shared libraries on HP-UX
1859 +) Fix for SHA1 assembly problem with MASM: it produces
1860 warnings about corrupt line number information when assembling
1861 with debugging information. This is caused by the overlapping
1863 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
1866 NCONF_get_number() has no error checking at all. As a replacement,
1867 NCONF_get_number_e() is defined (_e for "error checking") and is
1868 promoted strongly. The old NCONF_get_number is kept around for
1869 binary backward compatibility.
1870 Make it possible for methods to load from something other than a BIO,
1871 by providing a function pointer that is given a name instead of a BIO.
1872 For example, this could be used to load configuration data from an
1876 *) Fix typo in get_cert_by_subject() in by_dir.c
1877 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
1879 *) Rework the system to generate shared libraries:
1881 - Make note of the expected extension for the shared libraries and
1882 if there is a need for symbolic links from for example libcrypto.so.0
1883 to libcrypto.so.0.9.7. There is extended info in Configure for
1886 - Make as few rebuilds of the shared libraries as possible.
1888 - Still avoid linking the OpenSSL programs with the shared libraries.
1890 - When installing, install the shared libraries separately from the
1894 +) Fix for non blocking accept BIOs. Added new I/O special reason
1895 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
1896 with non blocking I/O was not possible because no retry code was
1897 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
1901 +) Added the beginnings of Rijndael support.
1904 +) Fix for bug in DirectoryString mask setting. Add support for
1905 X509_NAME_print_ex() in 'req' and X509_print_ex() function
1906 to allow certificate printing to more controllable, additional
1907 'certopt' option to 'x509' to allow new printing options to be
1911 +) Clean old EAY MD5 hack from e_os.h.
1914 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
1916 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
1917 and not in SSL_clear because the latter is also used by the
1918 accept/connect functions; previously, the settings made by
1919 SSL_set_read_ahead would be lost during the handshake.
1920 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
1922 *) Correct util/mkdef.pl to be selective about disabled algorithms.
1923 Previously, it would create entries for disableed algorithms no
1927 *) Added several new manual pages for SSL_* function.
1930 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
1932 *) In ssl23_get_client_hello, generate an error message when faced
1933 with an initial SSL 3.0/TLS record that is too small to contain the
1934 first two bytes of the ClientHello message, i.e. client_version.
1935 (Note that this is a pathologic case that probably has never happened
1936 in real life.) The previous approach was to use the version number
1937 from the record header as a substitute; but our protocol choice
1938 should not depend on that one because it is not authenticated
1939 by the Finished messages.
1942 *) More robust randomness gathering functions for Windows.
1943 [Jeffrey Altman <jaltman@columbia.edu>]
1945 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
1946 not set then we don't setup the error code for issuer check errors
1947 to avoid possibly overwriting other errors which the callback does
1948 handle. If an application does set the flag then we assume it knows
1949 what it is doing and can handle the new informational codes
1953 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
1954 a general "ANY" type, as such it should be able to decode anything
1955 including tagged types. However it didn't check the class so it would
1956 wrongly interpret tagged types in the same way as their universal
1957 counterpart and unknown types were just rejected. Changed so that the
1958 tagged and unknown types are handled in the same way as a SEQUENCE:
1959 that is the encoding is stored intact. There is also a new type
1960 "V_ASN1_OTHER" which is used when the class is not universal, in this
1961 case we have no idea what the actual type is so we just lump them all
1965 *) On VMS, stdout may very well lead to a file that is written to
1966 in a record-oriented fashion. That means that every write() will
1967 write a separate record, which will be read separately by the
1968 programs trying to read from it. This can be very confusing.
1970 The solution is to put a BIO filter in the way that will buffer
1971 text until a linefeed is reached, and then write everything a
1972 line at a time, so every record written will be an actual line,
1973 not chunks of lines and not (usually doesn't happen, but I've
1974 seen it once) several lines in one record. BIO_f_linebuffer() is
1977 Currently, it's a VMS-only method, because that's where it has
1978 been tested well enough.
1981 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
1982 it can return incorrect results.
1983 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
1984 but it was in 0.9.6-beta[12].)
1987 *) Disable the check for content being present when verifying detached
1988 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
1989 include zero length content when signing messages.
1992 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
1993 BIO_ctrl (for BIO pairs).
1996 *) Add DSO method for VMS.
1999 *) Bug fix: Montgomery multiplication could produce results with the
2003 *) Add RPM specification openssl.spec and modify it to build three
2004 packages. The default package contains applications, application
2005 documentation and run-time libraries. The devel package contains
2006 include files, static libraries and function documentation. The
2007 doc package contains the contents of the doc directory. The original
2008 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
2011 *) Add a large number of documentation files for many SSL routines.
2012 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
2014 *) Add a configuration entry for Sony News 4.
2015 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
2017 *) Don't set the two most significant bits to one when generating a
2018 random number < q in the DSA library.
2021 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
2022 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
2023 the underlying transport is blocking) if a handshake took place.
2024 (The default behaviour is needed by applications such as s_client
2025 and s_server that use select() to determine when to use SSL_read;
2026 but for applications that know in advance when to expect data, it
2027 just makes things more complicated.)
2030 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
2034 *) Add a few more EBCDIC conditionals that make `req' and `x509'
2035 work better on such systems.
2036 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
2038 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
2039 Update PKCS12_parse() so it copies the friendlyName and the
2040 keyid to the certificates aux info.
2043 *) Fix bug in PKCS7_verify() which caused an infinite loop
2044 if there was more than one signature.
2045 [Sven Uszpelkat <su@celocom.de>]
2047 *) Major change in util/mkdef.pl to include extra information
2048 about each symbol, as well as presentig variables as well
2049 as functions. This change means that there's n more need
2050 to rebuild the .num files when some algorithms are excluded.
2053 *) Allow the verify time to be set by an application,
2054 rather than always using the current time.
2057 *) Phase 2 verify code reorganisation. The certificate
2058 verify code now looks up an issuer certificate by a
2059 number of criteria: subject name, authority key id
2060 and key usage. It also verifies self signed certificates
2061 by the same criteria. The main comparison function is
2062 X509_check_issued() which performs these checks.
2064 Lot of changes were necessary in order to support this
2065 without completely rewriting the lookup code.
2067 Authority and subject key identifier are now cached.
2069 The LHASH 'certs' is X509_STORE has now been replaced
2070 by a STACK_OF(X509_OBJECT). This is mainly because an
2071 LHASH can't store or retrieve multiple objects with
2072 the same hash value.
2074 As a result various functions (which were all internal
2075 use only) have changed to handle the new X509_STORE
2076 structure. This will break anything that messed round
2077 with X509_STORE internally.
2079 The functions X509_STORE_add_cert() now checks for an
2080 exact match, rather than just subject name.
2082 The X509_STORE API doesn't directly support the retrieval
2083 of multiple certificates matching a given criteria, however
2084 this can be worked round by performing a lookup first
2085 (which will fill the cache with candidate certificates)
2086 and then examining the cache for matches. This is probably
2087 the best we can do without throwing out X509_LOOKUP
2088 entirely (maybe later...).
2090 The X509_VERIFY_CTX structure has been enhanced considerably.
2092 All certificate lookup operations now go via a get_issuer()
2093 callback. Although this currently uses an X509_STORE it
2094 can be replaced by custom lookups. This is a simple way
2095 to bypass the X509_STORE hackery necessary to make this
2096 work and makes it possible to use more efficient techniques
2097 in future. A very simple version which uses a simple
2098 STACK for its trusted certificate store is also provided
2099 using X509_STORE_CTX_trusted_stack().
2101 The verify_cb() and verify() callbacks now have equivalents
2102 in the X509_STORE_CTX structure.
2104 X509_STORE_CTX also has a 'flags' field which can be used
2105 to customise the verify behaviour.
2108 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
2109 excludes S/MIME capabilities.
2112 *) When a certificate request is read in keep a copy of the
2113 original encoding of the signed data and use it when outputing
2114 again. Signatures then use the original encoding rather than
2115 a decoded, encoded version which may cause problems if the
2116 request is improperly encoded.
2119 *) For consistency with other BIO_puts implementations, call
2120 buffer_write(b, ...) directly in buffer_puts instead of calling
2123 In BIO_puts, increment b->num_write as in BIO_write.
2124 [Peter.Sylvester@EdelWeb.fr]
2126 *) Fix BN_mul_word for the case where the word is 0. (We have to use
2127 BN_zero, we may not return a BIGNUM with an array consisting of
2131 *) Avoid calling abort() from within the library when problems are
2132 detected, except if preprocessor symbols have been defined
2133 (such as REF_CHECK, BN_DEBUG etc.).
2136 *) New openssl application 'rsautl'. This utility can be
2137 used for low level RSA operations. DER public key
2138 BIO/fp routines also added.
2141 *) New Configure entry and patches for compiling on QNX 4.
2142 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
2144 *) A demo state-machine implementation was sponsored by
2145 Nuron (http://www.nuron.com/) and is now available in
2146 demos/state_machine.
2149 *) New options added to the 'dgst' utility for signature
2150 generation and verification.
2153 *) Unrecognized PKCS#7 content types are now handled via a
2154 catch all ASN1_TYPE structure. This allows unsupported
2155 types to be stored as a "blob" and an application can
2156 encode and decode it manually.
2159 *) Fix various signed/unsigned issues to make a_strex.c
2161 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
2163 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
2164 length if passed a buffer. ASN1_INTEGER_to_BN failed
2165 if passed a NULL BN and its argument was negative.
2166 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
2168 *) Modification to PKCS#7 encoding routines to output definite
2169 length encoding. Since currently the whole structures are in
2170 memory there's not real point in using indefinite length
2171 constructed encoding. However if OpenSSL is compiled with
2172 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
2175 *) Added BIO_vprintf() and BIO_vsnprintf().
2178 *) Added more prefixes to parse for in the the strings written
2179 through a logging bio, to cover all the levels that are available
2180 through syslog. The prefixes are now:
2182 PANIC, EMERG, EMR => LOG_EMERG
2183 ALERT, ALR => LOG_ALERT
2184 CRIT, CRI => LOG_CRIT
2185 ERROR, ERR => LOG_ERR
2186 WARNING, WARN, WAR => LOG_WARNING
2187 NOTICE, NOTE, NOT => LOG_NOTICE
2188 INFO, INF => LOG_INFO
2189 DEBUG, DBG => LOG_DEBUG
2191 and as before, if none of those prefixes are present at the
2192 beginning of the string, LOG_ERR is chosen.
2194 On Win32, the LOG_* levels are mapped according to this:
2196 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
2197 LOG_WARNING => EVENTLOG_WARNING_TYPE
2198 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
2202 *) Made it possible to reconfigure with just the configuration
2203 argument "reconf" or "reconfigure". The command line arguments
2204 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
2205 and are retrieved from there when reconfiguring.
2209 [Assar Westerlund <assar@sics.se>, Richard Levitte]
2211 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
2214 *) The obj_dat.pl script was messing up the sorting of object
2215 names. The reason was that it compared the quoted version
2216 of strings as a result "OCSP" > "OCSP Signing" because
2217 " > SPACE. Changed script to store unquoted versions of
2218 names and add quotes on output. It was also omitting some
2219 names from the lookup table if they were given a default
2220 value (that is if SN is missing it is given the same
2221 value as LN and vice versa), these are now added on the
2222 grounds that if an object has a name we should be able to
2223 look it up. Finally added warning output when duplicate
2224 short or long names are found.
2227 *) Changes needed for Tandem NSK.
2228 [Scott Uroff <scott@xypro.com>]
2230 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
2231 RSA_padding_check_SSLv23(), special padding was never detected
2232 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
2233 version rollback attacks was not effective.
2235 In s23_clnt.c, don't use special rollback-attack detection padding
2236 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
2237 client; similarly, in s23_srvr.c, don't do the rollback check if
2238 SSL 2.0 is the only protocol enabled in the server.
2241 *) Make it possible to get hexdumps of unprintable data with 'openssl
2242 asn1parse'. By implication, the functions ASN1_parse_dump() and
2243 BIO_dump_indent() are added.
2246 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
2247 these print out strings and name structures based on various
2248 flags including RFC2253 support and proper handling of
2249 multibyte characters. Added options to the 'x509' utility
2250 to allow the various flags to be set.
2253 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
2254 Also change the functions X509_cmp_current_time() and
2255 X509_gmtime_adj() work with an ASN1_TIME structure,
2256 this will enable certificates using GeneralizedTime in validity
2257 dates to be checked.
2260 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
2261 negative public key encodings) on by default,
2262 NO_NEG_PUBKEY_BUG can be set to disable it.
2265 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
2266 content octets. An i2c_ASN1_OBJECT is unnecessary because
2267 the encoding can be trivially obtained from the structure.
2270 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
2271 not read locks (CRYPTO_r_[un]lock).
2274 *) A first attempt at creating official support for shared
2275 libraries through configuration. I've kept it so the
2276 default is static libraries only, and the OpenSSL programs
2277 are always statically linked for now, but there are
2278 preparations for dynamic linking in place.
2279 This has been tested on Linux and Tru64.
2282 *) Randomness polling function for Win9x, as described in:
2283 Peter Gutmann, Software Generation of Practically Strong
2287 *) Fix so PRNG is seeded in req if using an already existing
2291 *) New options to smime application. -inform and -outform
2292 allow alternative formats for the S/MIME message including
2293 PEM and DER. The -content option allows the content to be
2294 specified separately. This should allow things like Netscape
2295 form signing output easier to verify.
2298 *) Fix the ASN1 encoding of tags using the 'long form'.
2301 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
2302 STRING types. These convert content octets to and from the
2303 underlying type. The actual tag and length octets are
2304 already assumed to have been read in and checked. These
2305 are needed because all other string types have virtually
2306 identical handling apart from the tag. By having versions
2307 of the ASN1 functions that just operate on content octets
2308 IMPLICIT tagging can be handled properly. It also allows
2309 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
2310 and ASN1_INTEGER are identical apart from the tag.
2313 *) Change the handling of OID objects as follows:
2315 - New object identifiers are inserted in objects.txt, following
2316 the syntax given in objects.README.
2317 - objects.pl is used to process obj_mac.num and create a new
2319 - obj_dat.pl is used to create a new obj_dat.h, using the data in
2322 This is currently kind of a hack, and the perl code in objects.pl
2323 isn't very elegant, but it works as I intended. The simplest way
2324 to check that it worked correctly is to look in obj_dat.h and
2325 check the array nid_objs and make sure the objects haven't moved
2326 around (this is important!). Additions are OK, as well as
2327 consistent name changes.
2330 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
2333 *) Addition of the command line parameter '-rand file' to 'openssl req'.
2334 The given file adds to whatever has already been seeded into the
2335 random pool through the RANDFILE configuration file option or
2336 environment variable, or the default random state file.
2339 *) mkstack.pl now sorts each macro group into lexical order.
2340 Previously the output order depended on the order the files
2341 appeared in the directory, resulting in needless rewriting
2345 *) Patches to make OpenSSL compile under Win32 again. Mostly
2346 work arounds for the VC++ problem that it treats func() as
2347 func(void). Also stripped out the parts of mkdef.pl that
2348 added extra typesafe functions: these no longer exist.
2351 *) Reorganisation of the stack code. The macros are now all
2352 collected in safestack.h . Each macro is defined in terms of
2353 a "stack macro" of the form SKM_<name>(type, a, b). The
2354 DEBUG_SAFESTACK is now handled in terms of function casts,
2355 this has the advantage of retaining type safety without the
2356 use of additional functions. If DEBUG_SAFESTACK is not defined
2357 then the non typesafe macros are used instead. Also modified the
2358 mkstack.pl script to handle the new form. Needs testing to see
2359 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
2360 the default if no major problems. Similar behaviour for ASN1_SET_OF
2361 and PKCS12_STACK_OF.
2364 *) When some versions of IIS use the 'NET' form of private key the
2365 key derivation algorithm is different. Normally MD5(password) is
2366 used as a 128 bit RC4 key. In the modified case
2367 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
2368 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
2369 as the old Netscape_RSA functions except they have an additional
2370 'sgckey' parameter which uses the modified algorithm. Also added
2371 an -sgckey command line option to the rsa utility. Thanks to
2372 Adrian Peck <bertie@ncipher.com> for posting details of the modified
2373 algorithm to openssl-dev.
2376 *) The evp_local.h macros were using 'c.##kname' which resulted in
2377 invalid expansion on some systems (SCO 5.0.5 for example).
2378 Corrected to 'c.kname'.
2379 [Phillip Porch <root@theporch.com>]
2381 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
2382 a STACK of email addresses from a certificate or request, these look
2383 in the subject name and the subject alternative name extensions and
2384 omit any duplicate addresses.
2387 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
2388 This makes DSA verification about 2 % faster.
2391 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
2392 (meaning that now 2^5 values will be precomputed, which is only 4 KB
2393 plus overhead for 1024 bit moduli).
2394 This makes exponentiations about 0.5 % faster for 1024 bit
2395 exponents (as measured by "openssl speed rsa2048").
2398 *) Rename memory handling macros to avoid conflicts with other
2400 Malloc => OPENSSL_malloc
2401 Malloc_locked => OPENSSL_malloc_locked
2402 Realloc => OPENSSL_realloc
2403 Free => OPENSSL_free
2406 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
2407 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
2410 *) CygWin32 support.
2411 [John Jarvie <jjarvie@newsguy.com>]
2413 *) The type-safe stack code has been rejigged. It is now only compiled
2414 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
2415 by default all type-specific stack functions are "#define"d back to
2416 standard stack functions. This results in more streamlined output
2417 but retains the type-safety checking possibilities of the original
2421 *) The STACK code has been cleaned up, and certain type declarations
2422 that didn't make a lot of sense have been brought in line. This has
2423 also involved a cleanup of sorts in safestack.h to more correctly
2424 map type-safe stack functions onto their plain stack counterparts.
2425 This work has also resulted in a variety of "const"ifications of
2426 lots of the code, especially "_cmp" operations which should normally
2427 be prototyped with "const" parameters anyway.
2430 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
2431 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
2432 (The PRNG state consists of two parts, the large pool 'state' and 'md',
2433 where all of 'md' is used each time the PRNG is used, but 'state'
2434 is used only indexed by a cyclic counter. As entropy may not be
2435 well distributed from the beginning, 'md' is important as a
2436 chaining variable. However, the output function chains only half
2437 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
2438 all of 'md', and seeding with STATE_SIZE dummy bytes will result
2439 in all of 'state' being rewritten, with the new values depending
2440 on virtually all of 'md'. This overcomes the 80 bit limitation.)
2443 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
2444 the handshake is continued after ssl_verify_cert_chain();
2445 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
2446 can lead to 'unexplainable' connection aborts later.
2447 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
2449 *) Major EVP API cipher revision.
2450 Add hooks for extra EVP features. This allows various cipher
2451 parameters to be set in the EVP interface. Support added for variable
2452 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
2453 setting of RC2 and RC5 parameters.
2455 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
2458 Remove lots of duplicated code from the EVP library. For example *every*
2459 cipher init() function handles the 'iv' in the same way according to the
2460 cipher mode. They also all do nothing if the 'key' parameter is NULL and
2461 for CFB and OFB modes they zero ctx->num.
2463 New functionality allows removal of S/MIME code RC2 hack.
2465 Most of the routines have the same form and so can be declared in terms
2468 By shifting this to the top level EVP_CipherInit() it can be removed from
2469 all individual ciphers. If the cipher wants to handle IVs or keys
2470 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
2473 Change lots of functions like EVP_EncryptUpdate() to now return a
2474 value: although software versions of the algorithms cannot fail
2475 any installed hardware versions can.
2478 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
2479 this option is set, tolerate broken clients that send the negotiated
2480 protocol version number instead of the requested protocol version
2484 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
2485 i.e. non-zero for export ciphersuites, zero otherwise.
2486 Previous versions had this flag inverted, inconsistent with
2487 rsa_tmp_cb (..._TMP_RSA_CB).
2488 [Bodo Moeller; problem reported by Amit Chopra]
2490 *) Add missing DSA library text string. Work around for some IIS
2491 key files with invalid SEQUENCE encoding.
2494 *) Add a document (doc/standards.txt) that list all kinds of standards
2495 and so on that are implemented in OpenSSL.
2498 *) Enhance c_rehash script. Old version would mishandle certificates
2499 with the same subject name hash and wouldn't handle CRLs at all.
2500 Added -fingerprint option to crl utility, to support new c_rehash
2504 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
2507 *) Fix for SSL server purpose checking. Server checking was
2508 rejecting certificates which had extended key usage present
2509 but no ssl client purpose.
2510 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
2512 *) Make PKCS#12 code work with no password. The PKCS#12 spec
2513 is a little unclear about how a blank password is handled.
2514 Since the password in encoded as a BMPString with terminating
2515 double NULL a zero length password would end up as just the
2516 double NULL. However no password at all is different and is
2517 handled differently in the PKCS#12 key generation code. NS
2518 treats a blank password as zero length. MSIE treats it as no
2519 password on export: but it will try both on import. We now do
2520 the same: PKCS12_parse() tries zero length and no password if
2521 the password is set to "" or NULL (NULL is now a valid password:
2522 it wasn't before) as does the pkcs12 application.
2525 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
2526 perror when PEM_read_bio_X509_REQ fails, the error message must
2527 be obtained from the error queue.
2530 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
2531 it in ERR_remove_state if appropriate, and change ERR_get_state
2532 accordingly to avoid race conditions (this is necessary because
2533 thread_hash is no longer constant once set).
2536 *) Bugfix for linux-elf makefile.one.
2539 *) RSA_get_default_method() will now cause a default
2540 RSA_METHOD to be chosen if one doesn't exist already.
2541 Previously this was only set during a call to RSA_new()
2542 or RSA_new_method(NULL) meaning it was possible for
2543 RSA_get_default_method() to return NULL.
2546 *) Added native name translation to the existing DSO code
2547 that will convert (if the flag to do so is set) filenames
2548 that are sufficiently small and have no path information
2549 into a canonical native form. Eg. "blah" converted to
2550 "libblah.so" or "blah.dll" etc.
2553 *) New function ERR_error_string_n(e, buf, len) which is like
2554 ERR_error_string(e, buf), but writes at most 'len' bytes
2555 including the 0 terminator. For ERR_error_string_n, 'buf'
2557 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
2559 *) CONF library reworked to become more general. A new CONF
2560 configuration file reader "class" is implemented as well as a
2561 new functions (NCONF_*, for "New CONF") to handle it. The now
2562 old CONF_* functions are still there, but are reimplemented to
2563 work in terms of the new functions. Also, a set of functions
2564 to handle the internal storage of the configuration data is
2565 provided to make it easier to write new configuration file
2566 reader "classes" (I can definitely see something reading a
2567 configuration file in XML format, for example), called _CONF_*,
2568 or "the configuration storage API"...
2570 The new configuration file reading functions are:
2572 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
2573 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
2575 NCONF_default, NCONF_WIN32
2577 NCONF_dump_fp, NCONF_dump_bio
2579 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
2580 NCONF_new creates a new CONF object. This works in the same way
2581 as other interfaces in OpenSSL, like the BIO interface.
2582 NCONF_dump_* dump the internal storage of the configuration file,
2583 which is useful for debugging. All other functions take the same
2584 arguments as the old CONF_* functions wth the exception of the
2585 first that must be a `CONF *' instead of a `LHASH *'.
2587 To make it easer to use the new classes with the old CONF_* functions,
2588 the function CONF_set_default_method is provided.
2591 *) Add '-tls1' option to 'openssl ciphers', which was already
2592 mentioned in the documentation but had not been implemented.
2593 (This option is not yet really useful because even the additional
2594 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
2597 *) Initial DSO code added into libcrypto for letting OpenSSL (and
2598 OpenSSL-based applications) load shared libraries and bind to
2599 them in a portable way.
2600 [Geoff Thorpe, with contributions from Richard Levitte]
2602 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
2604 *) Make sure _lrotl and _lrotr are only used with MSVC.
2606 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
2607 (the default implementation of RAND_status).
2609 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
2610 to '-clrext' (= clear extensions), as intended and documented.
2611 [Bodo Moeller; inconsistency pointed out by Michael Attili
2614 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
2615 was larger than the MD block size.
2616 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
2618 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
2619 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
2620 using the passed key: if the passed key was a private key the result
2621 of X509_print(), for example, would be to print out all the private key
2625 *) des_quad_cksum() byte order bug fix.
2626 [Ulf Möller, using the problem description in krb4-0.9.7, where
2627 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
2629 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
2631 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
2633 *) For easily testing in shell scripts whether some command
2634 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
2635 returns with exit code 0 iff no command of the given name is available.
2636 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
2637 the output goes to stdout and nothing is printed to stderr.
2638 Additional arguments are always ignored.
2640 Since for each cipher there is a command of the same name,
2641 the 'no-cipher' compilation switches can be tested this way.
2643 ('openssl no-XXX' is not able to detect pseudo-commands such
2644 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
2647 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
2650 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
2651 is set; it will be thrown away anyway because each handshake creates
2653 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
2654 to parameters -- in previous versions (since OpenSSL 0.9.3) the
2655 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
2656 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
2659 *) New s_client option -ign_eof: EOF at stdin is ignored, and
2660 'Q' and 'R' lose their special meanings (quit/renegotiate).
2661 This is part of what -quiet does; unlike -quiet, -ign_eof
2662 does not suppress any output.
2665 *) Add compatibility options to the purpose and trust code. The
2666 purpose X509_PURPOSE_ANY is "any purpose" which automatically
2667 accepts a certificate or CA, this was the previous behaviour,
2668 with all the associated security issues.
2670 X509_TRUST_COMPAT is the old trust behaviour: only and
2671 automatically trust self signed roots in certificate store. A
2672 new trust setting X509_TRUST_DEFAULT is used to specify that
2673 a purpose has no associated trust setting and it should instead
2674 use the value in the default purpose.
2677 *) Fix the PKCS#8 DSA private key code so it decodes keys again
2678 and fix a memory leak.
2681 *) In util/mkerr.pl (which implements 'make errors'), preserve
2682 reason strings from the previous version of the .c file, as
2683 the default to have only downcase letters (and digits) in
2684 automatically generated reasons codes is not always appropriate.
2687 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
2688 using strerror. Previously, ERR_reason_error_string() returned
2689 library names as reason strings for SYSerr; but SYSerr is a special
2690 case where small numbers are errno values, not library numbers.
2693 *) Add '-dsaparam' option to 'openssl dhparam' application. This
2694 converts DSA parameters into DH parameters. (When creating parameters,
2695 DSA_generate_parameters is used.)
2698 *) Include 'length' (recommended exponent length) in C code generated
2699 by 'openssl dhparam -C'.
2702 *) The second argument to set_label in perlasm was already being used
2703 so couldn't be used as a "file scope" flag. Moved to third argument
2707 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
2708 instead of RAND_bytes for encryption IVs and salts.
2711 *) Include RAND_status() into RAND_METHOD instead of implementing
2712 it only for md_rand.c Otherwise replacing the PRNG by calling
2713 RAND_set_rand_method would be impossible.
2716 *) Don't let DSA_generate_key() enter an infinite loop if the random
2717 number generation fails.
2720 *) New 'rand' application for creating pseudo-random output.
2723 *) Added configuration support for Linux/IA64
2724 [Rolf Haberrecker <rolf@suse.de>]
2726 *) Assembler module support for Mingw32.
2729 *) Shared library support for HPUX (in shlib/).
2730 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
2732 *) Shared library support for Solaris gcc.
2733 [Lutz Behnke <behnke@trustcenter.de>]
2735 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
2737 *) PKCS7_encrypt() was adding text MIME headers twice because they
2738 were added manually and by SMIME_crlf_copy().
2741 *) In bntest.c don't call BN_rand with zero bits argument.
2742 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
2744 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
2745 case was implemented. This caused BN_div_recp() to fail occasionally.
2748 *) Add an optional second argument to the set_label() in the perl
2749 assembly language builder. If this argument exists and is set
2750 to 1 it signals that the assembler should use a symbol whose
2751 scope is the entire file, not just the current function. This
2752 is needed with MASM which uses the format label:: for this scope.
2753 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
2755 *) Change the ASN1 types so they are typedefs by default. Before
2756 almost all types were #define'd to ASN1_STRING which was causing
2757 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
2761 *) Change names of new functions to the new get1/get0 naming
2762 convention: After 'get1', the caller owns a reference count
2763 and has to call ..._free; 'get0' returns a pointer to some
2764 data structure without incrementing reference counters.
2765 (Some of the existing 'get' functions increment a reference
2766 counter, some don't.)
2767 Similarly, 'set1' and 'add1' functions increase reference
2768 counters or duplicate objects.
2771 *) Allow for the possibility of temp RSA key generation failure:
2772 the code used to assume it always worked and crashed on failure.
2775 *) Fix potential buffer overrun problem in BIO_printf().
2776 [Ulf Möller, using public domain code by Patrick Powell; problem
2777 pointed out by David Sacerdote <das33@cornell.edu>]
2779 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
2780 RAND_egd() and RAND_status(). In the command line application,
2781 the EGD socket can be specified like a seed file using RANDFILE
2785 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
2786 Some CAs (e.g. Verisign) distribute certificates in this form.
2789 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
2790 list to exclude them. This means that no special compilation option
2791 is needed to use anonymous DH: it just needs to be included in the
2795 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
2796 EVP_MD_type. The old functionality is available in a new macro called
2797 EVP_MD_md(). Change code that uses it and update docs.
2800 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
2801 where the 'void *' argument is replaced by a function pointer argument.
2802 Previously 'void *' was abused to point to functions, which works on
2803 many platforms, but is not correct. As these functions are usually
2804 called by macros defined in OpenSSL header files, most source code
2805 should work without changes.
2808 *) <openssl/opensslconf.h> (which is created by Configure) now contains
2809 sections with information on -D... compiler switches used for
2810 compiling the library so that applications can see them. To enable
2811 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
2812 must be defined. E.g.,
2813 #define OPENSSL_ALGORITHM_DEFINES
2814 #include <openssl/opensslconf.h>
2815 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
2816 [Richard Levitte, Ulf and Bodo Möller]
2818 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
2822 *) Change the 'other' type in certificate aux info to a STACK_OF
2823 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
2824 the required ASN1 format: arbitrary types determined by an OID.
2827 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
2828 argument to 'req'. This is not because the function is newer or
2829 better than others it just uses the work 'NEW' in the certificate
2830 request header lines. Some software needs this.
2833 *) Reorganise password command line arguments: now passwords can be
2834 obtained from various sources. Delete the PEM_cb function and make
2835 it the default behaviour: i.e. if the callback is NULL and the
2836 usrdata argument is not NULL interpret it as a null terminated pass
2837 phrase. If usrdata and the callback are NULL then the pass phrase
2838 is prompted for as usual.
2841 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
2842 the support is automatically enabled. The resulting binaries will
2843 autodetect the card and use it if present.
2844 [Ben Laurie and Compaq Inc.]
2846 *) Work around for Netscape hang bug. This sends certificate request
2847 and server done in one record. Since this is perfectly legal in the
2848 SSL/TLS protocol it isn't a "bug" option and is on by default. See
2849 the bugs/SSLv3 entry for more info.
2852 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
2855 *) Add -rand argument to smime and pkcs12 applications and read/write
2859 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
2862 *) Add command line password options to the remaining applications.
2865 *) Bug fix for BN_div_recp() for numerators with an even number of
2869 *) More tests in bntest.c, and changed test_bn output.
2872 *) ./config recognizes MacOS X now.
2875 *) Bug fix for BN_div() when the first words of num and divsor are
2876 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
2879 *) Add support for various broken PKCS#8 formats, and command line
2880 options to produce them.
2883 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
2884 get temporary BIGNUMs from a BN_CTX.
2887 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
2891 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
2892 include a #define from the old name to the new. The original intent
2893 was that statically linked binaries could for example just call
2894 SSLeay_add_all_ciphers() to just add ciphers to the table and not
2895 link with digests. This never worked becayse SSLeay_add_all_digests()
2896 and SSLeay_add_all_ciphers() were in the same source file so calling
2897 one would link with the other. They are now in separate source files.
2900 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
2903 *) Use a less unusual form of the Miller-Rabin primality test (it used
2904 a binary algorithm for exponentiation integrated into the Miller-Rabin
2905 loop, our standard modexp algorithms are faster).
2908 *) Support for the EBCDIC character set completed.
2909 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
2911 *) Source code cleanups: use const where appropriate, eliminate casts,
2912 use void * instead of char * in lhash.
2915 *) Bugfix: ssl3_send_server_key_exchange was not restartable
2916 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
2917 this the server could overwrite ephemeral keys that the client
2921 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
2922 using 50 iterations of the Rabin-Miller test.
2924 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
2925 iterations of the Rabin-Miller test as required by the appendix
2926 to FIPS PUB 186[-1]) instead of DSA_is_prime.
2927 As BN_is_prime_fasttest includes trial division, DSA parameter
2928 generation becomes much faster.
2930 This implies a change for the callback functions in DSA_is_prime
2931 and DSA_generate_parameters: The callback function is called once
2932 for each positive witness in the Rabin-Miller test, not just
2933 occasionally in the inner loop; and the parameters to the
2934 callback function now provide an iteration count for the outer
2935 loop rather than for the current invocation of the inner loop.
2936 DSA_generate_parameters additionally can call the callback
2937 function with an 'iteration count' of -1, meaning that a
2938 candidate has passed the trial division test (when q is generated
2939 from an application-provided seed, trial division is skipped).
2942 *) New function BN_is_prime_fasttest that optionally does trial
2943 division before starting the Rabin-Miller test and has
2944 an additional BN_CTX * argument (whereas BN_is_prime always
2945 has to allocate at least one BN_CTX).
2946 'callback(1, -1, cb_arg)' is called when a number has passed the
2947 trial division stage.
2950 *) Fix for bug in CRL encoding. The validity dates weren't being handled
2954 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
2957 *) New function BN_pseudo_rand().
2960 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
2961 bignum version of BN_from_montgomery() with the working code from
2962 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
2966 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
2967 made it impossible to use the same SSL_SESSION data structure in
2968 SSL2 clients in multiple threads.
2971 *) The return value of RAND_load_file() no longer counts bytes obtained
2972 by stat(). RAND_load_file(..., -1) is new and uses the complete file
2973 to seed the PRNG (previously an explicit byte count was required).
2974 [Ulf Möller, Bodo Möller]
2976 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
2977 used (char *) instead of (void *) and had casts all over the place.
2980 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
2983 *) Retain source code compatibility for BN_prime_checks macro:
2984 BN_is_prime(..., BN_prime_checks, ...) now uses
2985 BN_prime_checks_for_size to determine the appropriate number of
2986 Rabin-Miller iterations.
2989 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
2990 DH_CHECK_P_NOT_SAFE_PRIME.
2991 (Check if this is true? OpenPGP calls them "strong".)
2994 *) Merge the functionality of "dh" and "gendh" programs into a new program
2995 "dhparam". The old programs are retained for now but will handle DH keys
2996 (instead of parameters) in future.
2999 *) Make the ciphers, s_server and s_client programs check the return values
3000 when a new cipher list is set.
3003 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
3004 ciphers. Before when the 56bit ciphers were enabled the sorting was
3007 The syntax for the cipher sorting has been extended to support sorting by
3008 cipher-strength (using the strength_bits hard coded in the tables).
3009 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
3011 Fix a bug in the cipher-command parser: when supplying a cipher command
3012 string with an "undefined" symbol (neither command nor alphanumeric
3013 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
3014 an error is flagged.
3016 Due to the strength-sorting extension, the code of the
3017 ssl_create_cipher_list() function was completely rearranged. I hope that
3018 the readability was also increased :-)
3019 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
3021 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
3022 for the first serial number and places 2 in the serial number file. This
3023 avoids problems when the root CA is created with serial number zero and
3024 the first user certificate has the same issuer name and serial number
3028 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
3029 the new code. Add documentation for this stuff.
3032 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
3033 X509_*() to X509at_*() on the grounds that they don't handle X509
3034 structures and behave in an analagous way to the X509v3 functions:
3035 they shouldn't be called directly but wrapper functions should be used
3038 So we also now have some wrapper functions that call the X509at functions
3039 when passed certificate requests. (TO DO: similar things can be done with
3040 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
3041 things. Some of these need some d2i or i2d and print functionality
3042 because they handle more complex structures.)
3045 *) Add missing #ifndefs that caused missing symbols when building libssl
3046 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
3047 NO_RSA in ssl/s2*.c.
3048 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
3050 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
3051 has a return value which indicates the quality of the random data
3052 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
3053 error queue. New function RAND_pseudo_bytes() generates output that is
3054 guaranteed to be unique but not unpredictable. RAND_add is like
3055 RAND_seed, but takes an extra argument for an entropy estimate
3056 (RAND_seed always assumes full entropy).
3059 *) Do more iterations of Rabin-Miller probable prime test (specifically,
3060 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
3061 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
3062 in crypto/bn/bn_prime.c for the complete table). This guarantees a
3063 false-positive rate of at most 2^-80 for random input.
3066 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
3069 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
3070 in the 0.9.5 release), this returns the chain
3071 from an X509_CTX structure with a dup of the stack and all
3072 the X509 reference counts upped: so the stack will exist
3073 after X509_CTX_cleanup() has been called. Modify pkcs12.c
3076 Also make SSL_SESSION_print() print out the verify return
3080 *) Add manpage for the pkcs12 command. Also change the default
3081 behaviour so MAC iteration counts are used unless the new
3082 -nomaciter option is used. This improves file security and
3083 only older versions of MSIE (4.0 for example) need it.
3086 *) Honor the no-xxx Configure options when creating .DEF files.
3089 *) Add PKCS#10 attributes to field table: challengePassword,
3090 unstructuredName and unstructuredAddress. These are taken from
3091 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
3092 international characters are used.
3094 More changes to X509_ATTRIBUTE code: allow the setting of types
3095 based on strings. Remove the 'loc' parameter when adding
3096 attributes because these will be a SET OF encoding which is sorted
3100 *) Initial changes to the 'req' utility to allow request generation
3101 automation. This will allow an application to just generate a template
3102 file containing all the field values and have req construct the
3105 Initial support for X509_ATTRIBUTE handling. Stacks of these are
3106 used all over the place including certificate requests and PKCS#7
3107 structures. They are currently handled manually where necessary with
3108 some primitive wrappers for PKCS#7. The new functions behave in a
3109 manner analogous to the X509 extension functions: they allow
3110 attributes to be looked up by NID and added.
3112 Later something similar to the X509V3 code would be desirable to
3113 automatically handle the encoding, decoding and printing of the
3114 more complex types. The string types like challengePassword can
3115 be handled by the string table functions.
3117 Also modified the multi byte string table handling. Now there is
3118 a 'global mask' which masks out certain types. The table itself
3119 can use the flag STABLE_NO_MASK to ignore the mask setting: this
3120 is useful when for example there is only one permissible type
3121 (as in countryName) and using the mask might result in no valid
3125 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
3126 SSL_get_peer_finished to allow applications to obtain the latest
3127 Finished messages sent to the peer or expected from the peer,
3128 respectively. (SSL_get_peer_finished is usually the Finished message
3129 actually received from the peer, otherwise the protocol will be aborted.)
3131 As the Finished message are message digests of the complete handshake
3132 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
3133 be used for external authentication procedures when the authentication
3134 provided by SSL/TLS is not desired or is not enough.
3137 *) Enhanced support for Alpha Linux is added. Now ./config checks if
3138 the host supports BWX extension and if Compaq C is present on the
3139 $PATH. Just exploiting of the BWX extension results in 20-30%
3140 performance kick for some algorithms, e.g. DES and RC4 to mention
3141 a couple. Compaq C in turn generates ~20% faster code for MD5 and
3145 *) Add support for MS "fast SGC". This is arguably a violation of the
3146 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
3147 weak crypto and after checking the certificate is SGC a second one
3148 with strong crypto. MS SGC stops the first handshake after receiving
3149 the server certificate message and sends a second client hello. Since
3150 a server will typically do all the time consuming operations before
3151 expecting any further messages from the client (server key exchange
3152 is the most expensive) there is little difference between the two.
3154 To get OpenSSL to support MS SGC we have to permit a second client
3155 hello message after we have sent server done. In addition we have to
3156 reset the MAC if we do get this second client hello.
3159 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
3160 if a DER encoded private key is RSA or DSA traditional format. Changed
3161 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
3162 format DER encoded private key. Newer code should use PKCS#8 format which
3163 has the key type encoded in the ASN1 structure. Added DER private key
3164 support to pkcs8 application.
3167 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
3168 ciphersuites has been selected (as required by the SSL 3/TLS 1
3169 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
3170 is set, we interpret this as a request to violate the specification
3171 (the worst that can happen is a handshake failure, and 'correct'
3172 behaviour would result in a handshake failure anyway).
3175 *) In SSL_CTX_add_session, take into account that there might be multiple
3176 SSL_SESSION structures with the same session ID (e.g. when two threads
3177 concurrently obtain them from an external cache).
3178 The internal cache can handle only one SSL_SESSION with a given ID,
3179 so if there's a conflict, we now throw out the old one to achieve
3183 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
3184 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
3185 some routines that use cipher OIDs: some ciphers do not have OIDs
3186 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
3190 *) Simplify the trust setting structure and code. Now we just have
3191 two sequences of OIDs for trusted and rejected settings. These will
3192 typically have values the same as the extended key usage extension
3193 and any application specific purposes.
3195 The trust checking code now has a default behaviour: it will just
3196 check for an object with the same NID as the passed id. Functions can
3197 be provided to override either the default behaviour or the behaviour
3198 for a given id. SSL client, server and email already have functions
3199 in place for compatibility: they check the NID and also return "trusted"
3200 if the certificate is self signed.
3203 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
3204 traditional format into an EVP_PKEY structure.
3207 *) Add a password callback function PEM_cb() which either prompts for
3208 a password if usr_data is NULL or otherwise assumes it is a null
3209 terminated password. Allow passwords to be passed on command line
3210 environment or config files in a few more utilities.
3213 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
3214 keys. Add some short names for PKCS#8 PBE algorithms and allow them
3215 to be specified on the command line for the pkcs8 and pkcs12 utilities.
3216 Update documentation.
3219 *) Support for ASN1 "NULL" type. This could be handled before by using
3220 ASN1_TYPE but there wasn't any function that would try to read a NULL
3221 and produce an error if it couldn't. For compatibility we also have
3222 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
3223 don't allocate anything because they don't need to.
3226 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
3228 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
3230 *) Rebuild of the memory allocation routines used by OpenSSL code and
3231 possibly others as well. The purpose is to make an interface that
3232 provide hooks so anyone can build a separate set of allocation and
3233 deallocation routines to be used by OpenSSL, for example memory
3234 pool implementations, or something else, which was previously hard
3235 since Malloc(), Realloc() and Free() were defined as macros having
3236 the values malloc, realloc and free, respectively (except for Win32
3237 compilations). The same is provided for memory debugging code.
3238 OpenSSL already comes with functionality to find memory leaks, but
3239 this gives people a chance to debug other memory problems.
3241 With these changes, a new set of functions and macros have appeared:
3243 CRYPTO_set_mem_debug_functions() [F]
3244 CRYPTO_get_mem_debug_functions() [F]
3245 CRYPTO_dbg_set_options() [F]
3246 CRYPTO_dbg_get_options() [F]
3247 CRYPTO_malloc_debug_init() [M]
3249 The memory debug functions are NULL by default, unless the library
3250 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
3251 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
3252 gives the standard debugging functions that come with OpenSSL) or
3253 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
3254 provided by the library user) must be used. When the standard
3255 debugging functions are used, CRYPTO_dbg_set_options can be used to
3256 request additional information:
3257 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
3258 the CRYPTO_MDEBUG_xxx macro when compiling the library.
3260 Also, things like CRYPTO_set_mem_functions will always give the
3261 expected result (the new set of functions is used for allocation
3262 and deallocation) at all times, regardless of platform and compiler
3265 To finish it up, some functions that were never use in any other
3266 way than through macros have a new API and new semantic:
3269 CRYPTO_dbg_realloc()
3272 All macros of value have retained their old syntax.
3273 [Richard Levitte and Bodo Moeller]
3275 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
3276 ordering of SMIMECapabilities wasn't in "strength order" and there
3277 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
3281 *) Some ASN1 types with illegal zero length encoding (INTEGER,
3282 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
3283 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
3285 *) Merge in my S/MIME library for OpenSSL. This provides a simple
3286 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
3287 functionality to handle multipart/signed properly) and a utility
3288 called 'smime' to call all this stuff. This is based on code I
3289 originally wrote for Celo who have kindly allowed it to be
3290 included in OpenSSL.
3293 *) Add variants des_set_key_checked and des_set_key_unchecked of
3294 des_set_key (aka des_key_sched). Global variable des_check_key
3295 decides which of these is called by des_set_key; this way
3296 des_check_key behaves as it always did, but applications and
3297 the library itself, which was buggy for des_check_key == 1,
3298 have a cleaner way to pick the version they need.
3301 *) New function PKCS12_newpass() which changes the password of a
3305 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
3306 dynamic mix. In both cases the ids can be used as an index into the
3307 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
3308 functions so they accept a list of the field values and the
3309 application doesn't need to directly manipulate the X509_TRUST
3313 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
3317 *) Modify the way the V3 extension code looks up extensions. This now
3318 works in a similar way to the object code: we have some "standard"
3319 extensions in a static table which is searched with OBJ_bsearch()
3320 and the application can add dynamic ones if needed. The file
3321 crypto/x509v3/ext_dat.h now has the info: this file needs to be
3322 updated whenever a new extension is added to the core code and kept
3323 in ext_nid order. There is a simple program 'tabtest.c' which checks
3324 this. New extensions are not added too often so this file can readily
3325 be maintained manually.
3327 There are two big advantages in doing things this way. The extensions
3328 can be looked up immediately and no longer need to be "added" using
3329 X509V3_add_standard_extensions(): this function now does nothing.
3330 [Side note: I get *lots* of email saying the extension code doesn't
3331 work because people forget to call this function]
3332 Also no dynamic allocation is done unless new extensions are added:
3333 so if we don't add custom extensions there is no need to call
3334 X509V3_EXT_cleanup().
3337 *) Modify enc utility's salting as follows: make salting the default. Add a
3338 magic header, so unsalted files fail gracefully instead of just decrypting
3339 to garbage. This is because not salting is a big security hole, so people
3340 should be discouraged from doing it.
3343 *) Fixes and enhancements to the 'x509' utility. It allowed a message
3344 digest to be passed on the command line but it only used this
3345 parameter when signing a certificate. Modified so all relevant
3346 operations are affected by the digest parameter including the
3347 -fingerprint and -x509toreq options. Also -x509toreq choked if a
3348 DSA key was used because it didn't fix the digest.
3351 *) Initial certificate chain verify code. Currently tests the untrusted
3352 certificates for consistency with the verify purpose (which is set
3353 when the X509_STORE_CTX structure is set up) and checks the pathlength.
3355 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
3356 this is because it will reject chains with invalid extensions whereas
3357 every previous version of OpenSSL and SSLeay made no checks at all.
3359 Trust code: checks the root CA for the relevant trust settings. Trust
3360 settings have an initial value consistent with the verify purpose: e.g.
3361 if the verify purpose is for SSL client use it expects the CA to be
3362 trusted for SSL client use. However the default value can be changed to
3363 permit custom trust settings: one example of this would be to only trust
3364 certificates from a specific "secure" set of CAs.
3366 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
3367 which should be used for version portability: especially since the
3368 verify structure is likely to change more often now.
3370 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
3371 to set them. If not set then assume SSL clients will verify SSL servers
3374 Two new options to the verify program: -untrusted allows a set of
3375 untrusted certificates to be passed in and -purpose which sets the
3376 intended purpose of the certificate. If a purpose is set then the
3377 new chain verify code is used to check extension consistency.
3380 *) Support for the authority information access extension.
3383 *) Modify RSA and DSA PEM read routines to transparently handle
3384 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
3385 public keys in a format compatible with certificate
3386 SubjectPublicKeyInfo structures. Unfortunately there were already
3387 functions called *_PublicKey_* which used various odd formats so
3388 these are retained for compatibility: however the DSA variants were
3389 never in a public release so they have been deleted. Changed dsa/rsa
3390 utilities to handle the new format: note no releases ever handled public
3391 keys so we should be OK.
3393 The primary motivation for this change is to avoid the same fiasco
3394 that dogs private keys: there are several incompatible private key
3395 formats some of which are standard and some OpenSSL specific and
3396 require various evil hacks to allow partial transparent handling and
3397 even then it doesn't work with DER formats. Given the option anything
3398 other than PKCS#8 should be dumped: but the other formats have to
3399 stay in the name of compatibility.
3401 With public keys and the benefit of hindsight one standard format
3402 is used which works with EVP_PKEY, RSA or DSA structures: though
3403 it clearly returns an error if you try to read the wrong kind of key.
3405 Added a -pubkey option to the 'x509' utility to output the public key.
3406 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
3407 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
3408 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
3409 that do the same as the EVP_PKEY_assign_*() except they up the
3410 reference count of the added key (they don't "swallow" the
3414 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
3415 CRLs would fail if the file contained no certificates or no CRLs:
3416 added a new function to read in both types and return the number
3417 read: this means that if none are read it will be an error. The
3418 DER versions of the certificate and CRL reader would always fail
3419 because it isn't possible to mix certificates and CRLs in DER format
3420 without choking one or the other routine. Changed this to just read
3421 a certificate: this is the best we can do. Also modified the code
3422 in apps/verify.c to take notice of return codes: it was previously
3423 attempting to read in certificates from NULL pointers and ignoring
3424 any errors: this is one reason why the cert and CRL reader seemed
3425 to work. It doesn't check return codes from the default certificate
3426 routines: these may well fail if the certificates aren't installed.
3429 *) Code to support otherName option in GeneralName.
3432 *) First update to verify code. Change the verify utility
3433 so it warns if it is passed a self signed certificate:
3434 for consistency with the normal behaviour. X509_verify
3435 has been modified to it will now verify a self signed
3436 certificate if *exactly* the same certificate appears
3437 in the store: it was previously impossible to trust a
3438 single self signed certificate. This means that:
3439 openssl verify ss.pem
3440 now gives a warning about a self signed certificate but
3441 openssl verify -CAfile ss.pem ss.pem
3445 *) For servers, store verify_result in SSL_SESSION data structure
3446 (and add it to external session representation).
3447 This is needed when client certificate verifications fails,
3448 but an application-provided verification callback (set by
3449 SSL_CTX_set_cert_verify_callback) allows accepting the session
3450 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
3451 but returns 1): When the session is reused, we have to set
3452 ssl->verify_result to the appropriate error code to avoid
3454 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
3456 *) Fix a bug in the new PKCS#7 code: it didn't consider the
3457 case in PKCS7_dataInit() where the signed PKCS7 structure
3458 didn't contain any existing data because it was being created.
3459 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
3461 *) Add a salt to the key derivation routines in enc.c. This
3462 forms the first 8 bytes of the encrypted file. Also add a
3463 -S option to allow a salt to be input on the command line.
3466 *) New function X509_cmp(). Oddly enough there wasn't a function
3467 to compare two certificates. We do this by working out the SHA1
3468 hash and comparing that. X509_cmp() will be needed by the trust
3472 *) SSL_get1_session() is like SSL_get_session(), but increments
3473 the reference count in the SSL_SESSION returned.
3474 [Geoff Thorpe <geoff@eu.c2.net>]
3476 *) Fix for 'req': it was adding a null to request attributes.
3477 Also change the X509_LOOKUP and X509_INFO code to handle
3478 certificate auxiliary information.
3481 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
3485 *) Add the possibility to add extra information to the memory leak
3486 detecting output, to form tracebacks, showing from where each
3487 allocation was originated: CRYPTO_push_info("constant string") adds
3488 the string plus current file name and line number to a per-thread
3489 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
3490 is like calling CYRPTO_pop_info() until the stack is empty.
3491 Also updated memory leak detection code to be multi-thread-safe.
3494 *) Add options -text and -noout to pkcs7 utility and delete the
3495 encryption options which never did anything. Update docs.
3498 *) Add options to some of the utilities to allow the pass phrase
3499 to be included on either the command line (not recommended on
3500 OSes like Unix) or read from the environment. Update the
3501 manpages and fix a few bugs.
3504 *) Add a few manpages for some of the openssl commands.
3507 *) Fix the -revoke option in ca. It was freeing up memory twice,
3508 leaking and not finding already revoked certificates.
3511 *) Extensive changes to support certificate auxiliary information.
3512 This involves the use of X509_CERT_AUX structure and X509_AUX
3513 functions. An X509_AUX function such as PEM_read_X509_AUX()
3514 can still read in a certificate file in the usual way but it
3515 will also read in any additional "auxiliary information". By
3516 doing things this way a fair degree of compatibility can be
3517 retained: existing certificates can have this information added
3518 using the new 'x509' options.
3520 Current auxiliary information includes an "alias" and some trust
3521 settings. The trust settings will ultimately be used in enhanced
3522 certificate chain verification routines: currently a certificate
3523 can only be trusted if it is self signed and then it is trusted
3527 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
3528 The problem was that one of the replacement routines had not been working
3529 since SSLeay releases. For now the offending routine has been replaced
3530 with non-optimised assembler. Even so, this now gives around 95%
3531 performance improvement for 1024 bit RSA signs.
3534 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
3535 handling. Most clients have the effective key size in bits equal to
3536 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
3537 A few however don't do this and instead use the size of the decrypted key
3538 to determine the RC2 key length and the AlgorithmIdentifier to determine
3539 the effective key length. In this case the effective key length can still
3540 be 40 bits but the key length can be 168 bits for example. This is fixed
3541 by manually forcing an RC2 key into the EVP_PKEY structure because the
3542 EVP code can't currently handle unusual RC2 key sizes: it always assumes
3543 the key length and effective key length are equal.
3546 *) Add a bunch of functions that should simplify the creation of
3547 X509_NAME structures. Now you should be able to do:
3548 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
3549 and have it automatically work out the correct field type and fill in
3550 the structures. The more adventurous can try:
3551 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
3552 and it will (hopefully) work out the correct multibyte encoding.
3555 *) Change the 'req' utility to use the new field handling and multibyte
3556 copy routines. Before the DN field creation was handled in an ad hoc
3557 way in req, ca, and x509 which was rather broken and didn't support
3558 BMPStrings or UTF8Strings. Since some software doesn't implement
3559 BMPStrings or UTF8Strings yet, they can be enabled using the config file
3560 using the dirstring_type option. See the new comment in the default
3561 openssl.cnf for more info.
3564 *) Make crypto/rand/md_rand.c more robust:
3565 - Assure unique random numbers after fork().
3566 - Make sure that concurrent threads access the global counter and
3567 md serializably so that we never lose entropy in them
3568 or use exactly the same state in multiple threads.
3569 Access to the large state is not always serializable because
3570 the additional locking could be a performance killer, and
3571 md should be large enough anyway.
3574 *) New file apps/app_rand.c with commonly needed functionality
3575 for handling the random seed file.
3577 Use the random seed file in some applications that previously did not:
3579 dsaparam -genkey (which also ignored its '-rand' option),
3582 x509 (when signing).
3583 Except on systems with /dev/urandom, it is crucial to have a random
3584 seed file at least for key creation, DSA signing, and for DH exchanges;
3585 for RSA signatures we could do without one.
3587 gendh and gendsa (unlike genrsa) used to read only the first byte
3588 of each file listed in the '-rand' option. The function as previously
3589 found in genrsa is now in app_rand.c and is used by all programs
3590 that support '-rand'.
3593 *) In RAND_write_file, use mode 0600 for creating files;
3594 don't just chmod when it may be too late.
3597 *) Report an error from X509_STORE_load_locations
3598 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
3601 *) New function ASN1_mbstring_copy() this copies a string in either
3602 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
3603 into an ASN1_STRING type. A mask of permissible types is passed
3604 and it chooses the "minimal" type to use or an error if not type
3608 *) Add function equivalents to the various macros in asn1.h. The old
3609 macros are retained with an M_ prefix. Code inside the library can
3610 use the M_ macros. External code (including the openssl utility)
3611 should *NOT* in order to be "shared library friendly".
3614 *) Add various functions that can check a certificate's extensions
3615 to see if it usable for various purposes such as SSL client,
3616 server or S/MIME and CAs of these types. This is currently
3617 VERY EXPERIMENTAL but will ultimately be used for certificate chain
3618 verification. Also added a -purpose flag to x509 utility to
3619 print out all the purposes.
3622 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
3626 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
3627 for, obtain and decode and extension and obtain its critical flag.
3628 This allows all the necessary extension code to be handled in a
3629 single function call.
3632 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
3633 platforms. See crypto/rc4/rc4_enc.c for further details.
3636 *) New -noout option to asn1parse. This causes no output to be produced
3637 its main use is when combined with -strparse and -out to extract data
3638 from a file (which may not be in ASN.1 format).
3641 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
3642 when producing the local key id.
3643 [Richard Levitte <levitte@stacken.kth.se>]
3645 *) New option -dhparam in s_server. This allows a DH parameter file to be
3646 stated explicitly. If it is not stated then it tries the first server
3647 certificate file. The previous behaviour hard coded the filename
3651 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
3652 a public key to be input or output. For example:
3653 openssl rsa -in key.pem -pubout -out pubkey.pem
3654 Also added necessary DSA public key functions to handle this.
3657 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
3658 in the message. This was handled by allowing
3659 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
3660 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
3662 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
3663 to the end of the strings whereas this didn't. This would cause problems
3664 if strings read with d2i_ASN1_bytes() were later modified.
3665 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
3667 *) Fix for base64 decode bug. When a base64 bio reads only one line of
3668 data and it contains EOF it will end up returning an error. This is
3669 caused by input 46 bytes long. The cause is due to the way base64
3670 BIOs find the start of base64 encoded data. They do this by trying a
3671 trial decode on each line until they find one that works. When they
3672 do a flag is set and it starts again knowing it can pass all the
3673 data directly through the decoder. Unfortunately it doesn't reset
3674 the context it uses. This means that if EOF is reached an attempt
3675 is made to pass two EOFs through the context and this causes the
3676 resulting error. This can also cause other problems as well. As is
3677 usual with these problems it takes *ages* to find and the fix is
3678 trivial: move one line.
3679 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
3681 *) Ugly workaround to get s_client and s_server working under Windows. The
3682 old code wouldn't work because it needed to select() on sockets and the
3683 tty (for keypresses and to see if data could be written). Win32 only
3684 supports select() on sockets so we select() with a 1s timeout on the
3685 sockets and then see if any characters are waiting to be read, if none
3686 are present then we retry, we also assume we can always write data to
3687 the tty. This isn't nice because the code then blocks until we've
3688 received a complete line of data and it is effectively polling the
3689 keyboard at 1s intervals: however it's quite a bit better than not
3690 working at all :-) A dedicated Windows application might handle this
3691 with an event loop for example.
3694 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
3695 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
3696 will be called when RSA_sign() and RSA_verify() are used. This is useful
3697 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
3698 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
3699 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
3700 This necessitated the support of an extra signature type NID_md5_sha1
3701 for SSL signatures and modifications to the SSL library to use it instead
3702 of calling RSA_public_decrypt() and RSA_private_encrypt().
3705 *) Add new -verify -CAfile and -CApath options to the crl program, these
3706 will lookup a CRL issuers certificate and verify the signature in a
3707 similar way to the verify program. Tidy up the crl program so it
3708 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
3709 less strict. It will now permit CRL extensions even if it is not
3710 a V2 CRL: this will allow it to tolerate some broken CRLs.
3713 *) Initialize all non-automatic variables each time one of the openssl
3714 sub-programs is started (this is necessary as they may be started
3715 multiple times from the "OpenSSL>" prompt).
3716 [Lennart Bang, Bodo Moeller]
3718 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
3719 removing all other RSA functionality (this is what NO_RSA does). This
3720 is so (for example) those in the US can disable those operations covered
3721 by the RSA patent while allowing storage and parsing of RSA keys and RSA
3725 *) Non-copying interface to BIO pairs.
3726 (still largely untested)
3729 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
3730 ASCII string. This was handled independently in various places before.
3733 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
3734 UTF8 strings a character at a time.
3737 *) Use client_version from client hello to select the protocol
3738 (s23_srvr.c) and for RSA client key exchange verification
3739 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
3742 *) Add various utility functions to handle SPKACs, these were previously
3743 handled by poking round in the structure internals. Added new function
3744 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
3745 print, verify and generate SPKACs. Based on an original idea from
3746 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
3749 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
3752 *) Allow the config file extension section to be overwritten on the
3753 command line. Based on an original idea from Massimiliano Pala
3754 <madwolf@comune.modena.it>. The new option is called -extensions
3755 and can be applied to ca, req and x509. Also -reqexts to override
3756 the request extensions in req and -crlexts to override the crl extensions
3760 *) Add new feature to the SPKAC handling in ca. Now you can include
3761 the same field multiple times by preceding it by "XXXX." for example:
3764 this is the same syntax as used in the req config file.
3767 *) Allow certificate extensions to be added to certificate requests. These
3768 are specified in a 'req_extensions' option of the req section of the
3769 config file. They can be printed out with the -text option to req but
3770 are otherwise ignored at present.
3773 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
3774 data read consists of only the final block it would not decrypted because
3775 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
3776 A misplaced 'break' also meant the decrypted final block might not be
3777 copied until the next read.
3780 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
3781 a few extra parameters to the DH structure: these will be useful if
3782 for example we want the value of 'q' or implement X9.42 DH.
3785 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
3786 provides hooks that allow the default DSA functions or functions on a
3787 "per key" basis to be replaced. This allows hardware acceleration and
3788 hardware key storage to be handled without major modification to the
3789 library. Also added low level modexp hooks and CRYPTO_EX structure and
3790 associated functions.
3793 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
3794 as "read only": it can't be written to and the buffer it points to will
3795 not be freed. Reading from a read only BIO is much more efficient than
3796 a normal memory BIO. This was added because there are several times when
3797 an area of memory needs to be read from a BIO. The previous method was
3798 to create a memory BIO and write the data to it, this results in two
3799 copies of the data and an O(n^2) reading algorithm. There is a new
3800 function BIO_new_mem_buf() which creates a read only memory BIO from
3801 an area of memory. Also modified the PKCS#7 routines to use read only
3805 *) Bugfix: ssl23_get_client_hello did not work properly when called in
3806 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
3807 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
3808 but a retry condition occured while trying to read the rest.
3811 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
3812 NID_pkcs7_encrypted by default: this was wrong since this should almost
3813 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
3814 the encrypted data type: this is a more sensible place to put it and it
3815 allows the PKCS#12 code to be tidied up that duplicated this
3819 *) Changed obj_dat.pl script so it takes its input and output files on
3820 the command line. This should avoid shell escape redirection problems
3824 *) Initial support for certificate extension requests, these are included
3825 in things like Xenroll certificate requests. Included functions to allow
3826 extensions to be obtained and added.
3829 *) -crlf option to s_client and s_server for sending newlines as
3830 CRLF (as required by many protocols).
3833 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
3835 *) Install libRSAglue.a when OpenSSL is built with RSAref.
3836 [Ralf S. Engelschall]
3838 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
3839 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
3841 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
3845 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
3846 DH parameters/keys (q is lost during that conversion, but the resulting
3847 DH parameters contain its length).
3849 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
3850 much faster than DH_generate_parameters (which creates parameters
3851 where p = 2*q + 1), and also the smaller q makes DH computations
3852 much more efficient (160-bit exponentiation instead of 1024-bit
3853 exponentiation); so this provides a convenient way to support DHE
3854 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
3855 utter importance to use
3856 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
3858 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
3859 when such DH parameters are used, because otherwise small subgroup
3860 attacks may become possible!
3863 *) Avoid memory leak in i2d_DHparams.
3866 *) Allow the -k option to be used more than once in the enc program:
3867 this allows the same encrypted message to be read by multiple recipients.
3870 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
3871 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
3872 it will always use the numerical form of the OID, even if it has a short
3876 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
3877 method only got called if p,q,dmp1,dmq1,iqmp components were present,
3878 otherwise bn_mod_exp was called. In the case of hardware keys for example
3879 no private key components need be present and it might store extra data
3880 in the RSA structure, which cannot be accessed from bn_mod_exp.
3881 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
3882 private key operations.
3885 *) Added support for SPARC Linux.
3888 *) pem_password_cb function type incompatibly changed from
3889 typedef int pem_password_cb(char *buf, int size, int rwflag);
3891 ....(char *buf, int size, int rwflag, void *userdata);
3892 so that applications can pass data to their callbacks:
3893 The PEM[_ASN1]_{read,write}... functions and macros now take an
3894 additional void * argument, which is just handed through whenever
3895 the password callback is called.
3896 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
3898 New function SSL_CTX_set_default_passwd_cb_userdata.
3900 Compatibility note: As many C implementations push function arguments
3901 onto the stack in reverse order, the new library version is likely to
3902 interoperate with programs that have been compiled with the old
3903 pem_password_cb definition (PEM_whatever takes some data that
3904 happens to be on the stack as its last argument, and the callback
3905 just ignores this garbage); but there is no guarantee whatsoever that
3908 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
3909 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
3910 problems not only on Windows, but also on some Unix platforms.
3911 To avoid problematic command lines, these definitions are now in an
3912 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
3913 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
3916 *) MIPS III/IV assembler module is reimplemented.
3919 *) More DES library cleanups: remove references to srand/rand and
3920 delete an unused file.
3923 *) Add support for the the free Netwide assembler (NASM) under Win32,
3924 since not many people have MASM (ml) and it can be hard to obtain.
3925 This is currently experimental but it seems to work OK and pass all
3926 the tests. Check out INSTALL.W32 for info.
3929 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
3930 without temporary keys kept an extra copy of the server key,
3931 and connections with temporary keys did not free everything in case
3935 *) New function RSA_check_key and new openssl rsa option -check
3936 for verifying the consistency of RSA keys.
3937 [Ulf Moeller, Bodo Moeller]
3939 *) Various changes to make Win32 compile work:
3940 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
3941 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
3942 comparison" warnings.
3943 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
3946 *) Add a debugging option to PKCS#5 v2 key generation function: when
3947 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
3948 derived keys are printed to stderr.
3951 *) Copy the flags in ASN1_STRING_dup().
3952 [Roman E. Pavlov <pre@mo.msk.ru>]
3954 *) The x509 application mishandled signing requests containing DSA
3955 keys when the signing key was also DSA and the parameters didn't match.
3957 It was supposed to omit the parameters when they matched the signing key:
3958 the verifying software was then supposed to automatically use the CA's
3959 parameters if they were absent from the end user certificate.
3961 Omitting parameters is no longer recommended. The test was also
3962 the wrong way round! This was probably due to unusual behaviour in
3963 EVP_cmp_parameters() which returns 1 if the parameters match.
3964 This meant that parameters were omitted when they *didn't* match and
3965 the certificate was useless. Certificates signed with 'ca' didn't have
3967 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
3969 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
3970 The interface is as follows:
3971 Applications can use
3972 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
3973 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
3974 "off" is now the default.
3975 The library internally uses
3976 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
3977 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
3978 to disable memory-checking temporarily.
3980 Some inconsistent states that previously were possible (and were
3981 even the default) are now avoided.
3983 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
3984 with each memory chunk allocated; this is occasionally more helpful
3985 than just having a counter.
3987 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
3989 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
3993 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
3994 which largely parallels "options", but is for changing API behaviour,
3995 whereas "options" are about protocol behaviour.
3996 Initial "mode" flags are:
3998 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
3999 a single record has been written.
4000 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
4001 retries use the same buffer location.
4002 (But all of the contents must be
4006 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
4009 *) Fix problems with no-hmac etc.
4010 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
4012 *) New functions RSA_get_default_method(), RSA_set_method() and
4013 RSA_get_method(). These allows replacement of RSA_METHODs without having
4014 to mess around with the internals of an RSA structure.
4017 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
4018 Also really enable memory leak checks in openssl.c and in some
4020 [Chad C. Mulligan, Bodo Moeller]
4022 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
4023 up the length of negative integers. This has now been simplified to just
4024 store the length when it is first determined and use it later, rather
4025 than trying to keep track of where data is copied and updating it to
4027 [Steve Henson, reported by Brien Wheeler
4028 <bwheeler@authentica-security.com>]
4030 *) Add a new function PKCS7_signatureVerify. This allows the verification
4031 of a PKCS#7 signature but with the signing certificate passed to the
4032 function itself. This contrasts with PKCS7_dataVerify which assumes the
4033 certificate is present in the PKCS#7 structure. This isn't always the
4034 case: certificates can be omitted from a PKCS#7 structure and be
4035 distributed by "out of band" means (such as a certificate database).
4038 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
4039 function prototypes in pem.h, also change util/mkdef.pl to add the
4040 necessary function names.
4043 *) mk1mf.pl (used by Windows builds) did not properly read the
4044 options set by Configure in the top level Makefile, and Configure
4045 was not even able to write more than one option correctly.
4046 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
4049 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
4050 file to be loaded from a BIO or FILE pointer. The BIO version will
4051 for example allow memory BIOs to contain config info.
4054 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
4055 Whoever hopes to achieve shared-library compatibility across versions
4056 must use this, not the compile-time macro.
4057 (Exercise 0.9.4: Which is the minimum library version required by
4059 Note: All this applies only to multi-threaded programs, others don't
4063 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
4064 through a BIO pair triggered the default case, i.e.
4065 SSLerr(...,SSL_R_UNKNOWN_STATE).
4068 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
4069 can use the SSL library even if none of the specific BIOs is
4073 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
4074 for the encoded length.
4075 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
4077 *) Add initial documentation of the X509V3 functions.
4080 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
4081 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
4082 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
4083 secure PKCS#8 private key format with a high iteration count.
4086 *) Fix determination of Perl interpreter: A perl or perl5
4087 _directory_ in $PATH was also accepted as the interpreter.
4088 [Ralf S. Engelschall]
4090 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
4091 wrong with it but it was very old and did things like calling
4092 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
4096 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
4097 to use the new extension code.
4100 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
4101 with macros. This should make it easier to change their form, add extra
4102 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
4106 *) Add to configuration table a new entry that can specify an alternative
4107 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
4108 according to Mark Crispin <MRC@Panda.COM>.
4112 *) DES CBC did not update the IV. Weird.
4115 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
4116 Changing the behaviour of the former might break existing programs --
4117 where IV updating is needed, des_ncbc_encrypt can be used.
4120 *) When bntest is run from "make test" it drives bc to check its
4121 calculations, as well as internally checking them. If an internal check
4122 fails, it needs to cause bc to give a non-zero result or make test carries
4123 on without noticing the failure. Fixed.
4126 *) DES library cleanups.
4129 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
4130 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
4131 ciphers. NOTE: although the key derivation function has been verified
4132 against some published test vectors it has not been extensively tested
4133 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
4137 *) Instead of "mkdir -p", which is not fully portable, use new
4138 Perl script "util/mkdir-p.pl".
4141 *) Rewrite the way password based encryption (PBE) is handled. It used to
4142 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
4143 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
4144 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
4145 the 'parameter' field of the AlgorithmIdentifier is passed to the
4146 underlying key generation function so it must do its own ASN1 parsing.
4147 This has also changed the EVP_PBE_CipherInit() function which now has a
4148 'parameter' argument instead of literal salt and iteration count values
4149 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
4152 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
4153 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
4154 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
4155 KEY" because this clashed with PKCS#8 unencrypted string. Since this
4156 value was just used as a "magic string" and not used directly its
4157 value doesn't matter.
4160 *) Introduce some semblance of const correctness to BN. Shame C doesn't
4164 *) "linux-sparc64" configuration (ultrapenguin).
4165 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
4166 "linux-sparc" configuration.
4167 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
4169 *) config now generates no-xxx options for missing ciphers.
4172 *) Support the EBCDIC character set (work in progress).
4173 File ebcdic.c not yet included because it has a different license.
4174 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
4176 *) Support BS2000/OSD-POSIX.
4177 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
4179 *) Make callbacks for key generation use void * instead of char *.
4182 *) Make S/MIME samples compile (not yet tested).
4185 *) Additional typesafe stacks.
4188 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
4192 Changes between 0.9.3 and 0.9.3a [29 May 1999]
4194 *) New configuration variant "sco5-gcc".
4196 *) Updated some demos.
4197 [Sean O Riordain, Wade Scholine]
4199 *) Add missing BIO_free at exit of pkcs12 application.
4202 *) Fix memory leak in conf.c.
4205 *) Updates for Win32 to assembler version of MD5.
4208 *) Set #! path to perl in apps/der_chop to where we found it
4209 instead of using a fixed path.
4212 *) SHA library changes for irix64-mips4-cc.
4215 *) Improvements for VMS support.
4219 Changes between 0.9.2b and 0.9.3 [24 May 1999]
4221 *) Bignum library bug fix. IRIX 6 passes "make test" now!
4222 This also avoids the problems with SC4.2 and unpatched SC5.
4223 [Andy Polyakov <appro@fy.chalmers.se>]
4225 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
4226 These are required because of the typesafe stack would otherwise break
4227 existing code. If old code used a structure member which used to be STACK
4228 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
4229 sk_num or sk_value it would produce an error because the num, data members
4230 are not present in STACK_OF. Now it just produces a warning. sk_set
4231 replaces the old method of assigning a value to sk_value
4232 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
4233 that does this will no longer work (and should use sk_set instead) but
4234 this could be regarded as a "questionable" behaviour anyway.
4237 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
4238 correctly handle encrypted S/MIME data.
4241 *) Change type of various DES function arguments from des_cblock
4242 (which means, in function argument declarations, pointer to char)
4243 to des_cblock * (meaning pointer to array with 8 char elements),
4244 which allows the compiler to do more typechecking; it was like
4245 that back in SSLeay, but with lots of ugly casts.
4247 Introduce new type const_des_cblock.
4250 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
4251 problems: find RecipientInfo structure that matches recipient certificate
4252 and initialise the ASN1 structures properly based on passed cipher.
4255 *) Belatedly make the BN tests actually check the results.
4258 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
4259 to and from BNs: it was completely broken. New compilation option
4260 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
4261 key elements as negative integers.
4264 *) Reorganize and speed up MD5.
4265 [Andy Polyakov <appro@fy.chalmers.se>]
4268 [Richard Levitte <richard@levitte.org>]
4270 *) New option -out to asn1parse to allow the parsed structure to be
4271 output to a file. This is most useful when combined with the -strparse
4272 option to examine the output of things like OCTET STRINGS.
4275 *) Make SSL library a little more fool-proof by not requiring any longer
4276 that SSL_set_{accept,connect}_state be called before
4277 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
4278 in many applications because usually everything *appeared* to work as
4279 intended anyway -- now it really works as intended).
4282 *) Move openssl.cnf out of lib/.
4285 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
4286 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
4287 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
4288 [Ralf S. Engelschall]
4290 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
4291 handle PKCS#7 enveloped data properly.
4292 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
4294 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
4295 copying pointers. The cert_st handling is changed by this in
4296 various ways (and thus what used to be known as ctx->default_cert
4297 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
4298 any longer when s->cert does not give us what we need).
4299 ssl_cert_instantiate becomes obsolete by this change.
4300 As soon as we've got the new code right (possibly it already is?),
4301 we have solved a couple of bugs of the earlier code where s->cert
4302 was used as if it could not have been shared with other SSL structures.
4304 Note that using the SSL API in certain dirty ways now will result
4305 in different behaviour than observed with earlier library versions:
4306 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
4307 does not influence s as it used to.
4309 In order to clean up things more thoroughly, inside SSL_SESSION
4310 we don't use CERT any longer, but a new structure SESS_CERT
4311 that holds per-session data (if available); currently, this is
4312 the peer's certificate chain and, for clients, the server's certificate
4313 and temporary key. CERT holds only those values that can have
4314 meaningful defaults in an SSL_CTX.
4317 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
4318 from the internal representation. Various PKCS#7 fixes: remove some
4319 evil casts and set the enc_dig_alg field properly based on the signing
4323 *) Allow PKCS#12 password to be set from the command line or the
4324 environment. Let 'ca' get its config file name from the environment
4325 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
4329 *) Allow certificate policies extension to use an IA5STRING for the
4330 organization field. This is contrary to the PKIX definition but
4331 VeriSign uses it and IE5 only recognises this form. Document 'x509'
4335 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
4336 without disallowing inline assembler and the like for non-pedantic builds.
4339 *) Support Borland C++ builder.
4340 [Janez Jere <jj@void.si>, modified by Ulf Möller]
4345 *) SHA-1 cleanups and performance enhancements.
4346 [Andy Polyakov <appro@fy.chalmers.se>]
4348 *) Sparc v8plus assembler for the bignum library.
4349 [Andy Polyakov <appro@fy.chalmers.se>]
4351 *) Accept any -xxx and +xxx compiler options in Configure.
4354 *) Update HPUX configuration.
4357 *) Add missing sk_<type>_unshift() function to safestack.h
4358 [Ralf S. Engelschall]
4360 *) New function SSL_CTX_use_certificate_chain_file that sets the
4361 "extra_cert"s in addition to the certificate. (This makes sense
4362 only for "PEM" format files, as chains as a whole are not
4366 *) Support verify_depth from the SSL API.
4367 x509_vfy.c had what can be considered an off-by-one-error:
4368 Its depth (which was not part of the external interface)
4369 was actually counting the number of certificates in a chain;
4370 now it really counts the depth.
4373 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
4374 instead of X509err, which often resulted in confusing error
4375 messages since the error codes are not globally unique
4376 (e.g. an alleged error in ssl3_accept when a certificate
4377 didn't match the private key).
4379 *) New function SSL_CTX_set_session_id_context that allows to set a default
4380 value (so that you don't need SSL_set_session_id_context for each
4381 connection using the SSL_CTX).
4384 *) OAEP decoding bug fix.
4387 *) Support INSTALL_PREFIX for package builders, as proposed by
4391 *) New Configure options "threads" and "no-threads". For systems
4392 where the proper compiler options are known (currently Solaris
4393 and Linux), "threads" is the default.
4396 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
4399 *) Install various scripts to $(OPENSSLDIR)/misc, not to
4400 $(INSTALLTOP)/bin -- they shouldn't clutter directories
4401 such as /usr/local/bin.
4404 *) "make linux-shared" to build shared libraries.
4405 [Niels Poppe <niels@netbox.org>]
4407 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
4410 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
4411 extension adding in x509 utility.
4414 *) Remove NOPROTO sections and error code comments.
4417 *) Partial rewrite of the DEF file generator to now parse the ANSI
4421 *) New Configure options --prefix=DIR and --openssldir=DIR.
4424 *) Complete rewrite of the error code script(s). It is all now handled
4425 by one script at the top level which handles error code gathering,
4426 header rewriting and C source file generation. It should be much better
4427 than the old method: it now uses a modified version of Ulf's parser to
4428 read the ANSI prototypes in all header files (thus the old K&R definitions
4429 aren't needed for error creation any more) and do a better job of
4430 translating function codes into names. The old 'ASN1 error code imbedded
4431 in a comment' is no longer necessary and it doesn't use .err files which
4432 have now been deleted. Also the error code call doesn't have to appear all
4433 on one line (which resulted in some large lines...).
4436 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
4439 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
4440 0 (which usually indicates a closed connection), but continue reading.
4443 *) Fix some race conditions.
4446 *) Add support for CRL distribution points extension. Add Certificate
4447 Policies and CRL distribution points documentation.
4450 *) Move the autogenerated header file parts to crypto/opensslconf.h.
4453 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
4454 8 of keying material. Merlin has also confirmed interop with this fix
4455 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
4456 [Merlin Hughes <merlin@baltimore.ie>]
4458 *) Fix lots of warnings.
4459 [Richard Levitte <levitte@stacken.kth.se>]
4461 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
4462 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
4463 [Richard Levitte <levitte@stacken.kth.se>]
4465 *) Fix problems with sizeof(long) == 8.
4466 [Andy Polyakov <appro@fy.chalmers.se>]
4468 *) Change functions to ANSI C.
4471 *) Fix typos in error codes.
4472 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
4474 *) Remove defunct assembler files from Configure.
4477 *) SPARC v8 assembler BIGNUM implementation.
4478 [Andy Polyakov <appro@fy.chalmers.se>]
4480 *) Support for Certificate Policies extension: both print and set.
4481 Various additions to support the r2i method this uses.
4484 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
4485 return a const string when you are expecting an allocated buffer.
4488 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
4489 types DirectoryString and DisplayText.
4492 *) Add code to allow r2i extensions to access the configuration database,
4493 add an LHASH database driver and add several ctx helper functions.
4496 *) Fix an evil bug in bn_expand2() which caused various BN functions to
4497 fail when they extended the size of a BIGNUM.
4500 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
4501 support typesafe stack.
4504 *) Fix typo in SSL_[gs]et_options().
4505 [Nils Frostberg <nils@medcom.se>]
4507 *) Delete various functions and files that belonged to the (now obsolete)
4508 old X509V3 handling code.
4511 *) New Configure option "rsaref".
4514 *) Don't auto-generate pem.h.
4517 *) Introduce type-safe ASN.1 SETs.
4520 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
4521 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
4523 *) Introduce type-safe STACKs. This will almost certainly break lots of code
4524 that links with OpenSSL (well at least cause lots of warnings), but fear
4525 not: the conversion is trivial, and it eliminates loads of evil casts. A
4526 few STACKed things have been converted already. Feel free to convert more.
4527 In the fullness of time, I'll do away with the STACK type altogether.
4530 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
4531 specified in <certfile> by updating the entry in the index.txt file.
4532 This way one no longer has to edit the index.txt file manually for
4533 revoking a certificate. The -revoke option does the gory details now.
4534 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
4536 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
4537 `-text' option at all and this way the `-noout -text' combination was
4538 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
4539 [Ralf S. Engelschall]
4541 *) Make sure a corresponding plain text error message exists for the
4542 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
4543 verify callback function determined that a certificate was revoked.
4544 [Ralf S. Engelschall]
4546 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
4547 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
4548 all available cipers including rc5, which was forgotten until now.
4549 In order to let the testing shell script know which algorithms
4550 are available, a new (up to now undocumented) command
4551 "openssl list-cipher-commands" is used.
4554 *) Bugfix: s_client occasionally would sleep in select() when
4555 it should have checked SSL_pending() first.
4558 *) New functions DSA_do_sign and DSA_do_verify to provide access to
4559 the raw DSA values prior to ASN.1 encoding.
4562 *) Tweaks to Configure
4563 [Niels Poppe <niels@netbox.org>]
4565 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
4569 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
4572 *) New config option to avoid instructions that are illegal on the 80386.
4573 The default code is faster, but requires at least a 486.
4576 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
4577 SSL2_SERVER_VERSION (not used at all) macros, which are now the
4578 same as SSL2_VERSION anyway.
4581 *) New "-showcerts" option for s_client.
4584 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
4585 application. Various cleanups and fixes.
4588 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
4589 modify error routines to work internally. Add error codes and PBE init
4590 to library startup routines.
4593 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
4594 packing functions to asn1 and evp. Changed function names and error
4595 codes along the way.
4598 *) PKCS12 integration: and so it begins... First of several patches to
4599 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
4600 objects to objects.h
4603 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
4604 and display support for Thawte strong extranet extension.
4607 *) Add LinuxPPC support.
4608 [Jeff Dubrule <igor@pobox.org>]
4610 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
4611 bn_div_words in alpha.s.
4612 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
4614 *) Make sure the RSA OAEP test is skipped under -DRSAref because
4615 OAEP isn't supported when OpenSSL is built with RSAref.
4616 [Ulf Moeller <ulf@fitug.de>]
4618 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
4619 so they no longer are missing under -DNOPROTO.
4620 [Soren S. Jorvang <soren@t.dk>]
4623 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
4625 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
4626 doesn't work when the session is reused. Coming soon!
4629 *) Fix a security hole, that allows sessions to be reused in the wrong
4630 context thus bypassing client cert protection! All software that uses
4631 client certs and session caches in multiple contexts NEEDS PATCHING to
4632 allow session reuse! A fuller solution is in the works.
4633 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
4635 *) Some more source tree cleanups (removed obsolete files
4636 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
4637 permission on "config" script to be executable) and a fix for the INSTALL
4639 [Ulf Moeller <ulf@fitug.de>]
4641 *) Remove some legacy and erroneous uses of malloc, free instead of
4643 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
4645 *) Make rsa_oaep_test return non-zero on error.
4646 [Ulf Moeller <ulf@fitug.de>]
4648 *) Add support for native Solaris shared libraries. Configure
4649 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
4650 if someone would make that last step automatic.
4651 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
4653 *) ctx_size was not built with the right compiler during "make links". Fixed.
4656 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
4657 except NULL ciphers". This means the default cipher list will no longer
4658 enable NULL ciphers. They need to be specifically enabled e.g. with
4659 the string "DEFAULT:eNULL".
4662 *) Fix to RSA private encryption routines: if p < q then it would
4663 occasionally produce an invalid result. This will only happen with
4664 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
4667 *) Be less restrictive and allow also `perl util/perlpath.pl
4668 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
4669 because this way one can also use an interpreter named `perl5' (which is
4670 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
4671 installed as `perl').
4672 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
4674 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
4675 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
4677 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
4678 advapi32.lib to Win32 build and change the pem test comparision
4679 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
4680 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
4681 and crypto/des/ede_cbcm_enc.c.
4684 *) DES quad checksum was broken on big-endian architectures. Fixed.
4687 *) Comment out two functions in bio.h that aren't implemented. Fix up the
4688 Win32 test batch file so it (might) work again. The Win32 test batch file
4689 is horrible: I feel ill....
4692 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
4693 in e_os.h. Audit of header files to check ANSI and non ANSI
4694 sections: 10 functions were absent from non ANSI section and not exported
4695 from Windows DLLs. Fixed up libeay.num for new functions.
4698 *) Make `openssl version' output lines consistent.
4699 [Ralf S. Engelschall]
4701 *) Fix Win32 symbol export lists for BIO functions: Added
4702 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
4703 to ms/libeay{16,32}.def.
4704 [Ralf S. Engelschall]
4706 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
4707 fine under Unix and passes some trivial tests I've now added. But the
4708 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
4709 added to make sure no one expects that this stuff really works in the
4710 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
4711 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
4713 [Ralf S. Engelschall]
4715 *) Fix the generation of two part addresses in perl.
4716 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
4718 *) Add config entry for Linux on MIPS.
4719 [John Tobey <jtobey@channel1.com>]
4721 *) Make links whenever Configure is run, unless we are on Windoze.
4724 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
4725 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
4729 *) Add a useful kludge to allow package maintainers to specify compiler and
4730 other platforms details on the command line without having to patch the
4731 Configure script everytime: One now can use ``perl Configure
4732 <id>:<details>'', i.e. platform ids are allowed to have details appended
4733 to them (seperated by colons). This is treated as there would be a static
4734 pre-configured entry in Configure's %table under key <id> with value
4735 <details> and ``perl Configure <id>'' is called. So, when you want to
4736 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
4737 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
4738 now, which overrides the FreeBSD-elf entry on-the-fly.
4739 [Ralf S. Engelschall]
4741 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
4744 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
4745 on the `perl Configure ...' command line. This way one can compile
4746 OpenSSL libraries with Position Independent Code (PIC) which is needed
4747 for linking it into DSOs.
4748 [Ralf S. Engelschall]
4750 *) Remarkably, export ciphers were totally broken and no-one had noticed!
4754 *) Cleaned up the LICENSE document: The official contact for any license
4755 questions now is the OpenSSL core team under openssl-core@openssl.org.
4756 And add a paragraph about the dual-license situation to make sure people
4757 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
4758 to the OpenSSL toolkit.
4759 [Ralf S. Engelschall]
4761 *) General source tree makefile cleanups: Made `making xxx in yyy...'
4762 display consistent in the source tree and replaced `/bin/rm' by `rm'.
4763 Additonally cleaned up the `make links' target: Remove unnecessary
4764 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
4765 to speed processing and no longer clutter the display with confusing
4766 stuff. Instead only the actually done links are displayed.
4767 [Ralf S. Engelschall]
4769 *) Permit null encryption ciphersuites, used for authentication only. It used
4770 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
4771 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
4775 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
4776 signed attributes when verifying signatures (this would break them),
4777 the detached data encoding was wrong and public keys obtained using
4778 X509_get_pubkey() weren't freed.
4781 *) Add text documentation for the BUFFER functions. Also added a work around
4782 to a Win95 console bug. This was triggered by the password read stuff: the
4783 last character typed gets carried over to the next fread(). If you were
4784 generating a new cert request using 'req' for example then the last
4785 character of the passphrase would be CR which would then enter the first
4789 *) Added the new `Includes OpenSSL Cryptography Software' button as
4790 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
4791 button and can be used by applications based on OpenSSL to show the
4792 relationship to the OpenSSL project.
4793 [Ralf S. Engelschall]
4795 *) Remove confusing variables in function signatures in files
4796 ssl/ssl_lib.c and ssl/ssl.h.
4797 [Lennart Bong <lob@kulthea.stacken.kth.se>]
4799 *) Don't install bss_file.c under PREFIX/include/
4800 [Lennart Bong <lob@kulthea.stacken.kth.se>]
4802 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
4803 functions that return function pointers and has support for NT specific
4804 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
4805 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
4806 unsigned to signed types: this was killing the Win32 compile.
4809 *) Add new certificate file to stack functions,
4810 SSL_add_dir_cert_subjects_to_stack() and
4811 SSL_add_file_cert_subjects_to_stack(). These largely supplant
4812 SSL_load_client_CA_file(), and can be used to add multiple certs easily
4813 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
4814 This means that Apache-SSL and similar packages don't have to mess around
4815 to add as many CAs as they want to the preferred list.
4818 *) Experiment with doxygen documentation. Currently only partially applied to
4820 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
4821 openssl.doxy as the configuration file.
4824 *) Get rid of remaining C++-style comments which strict C compilers hate.
4825 [Ralf S. Engelschall, pointed out by Carlos Amengual]
4827 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
4828 compiled in by default: it has problems with large keys.
4831 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
4832 DH private keys and/or callback functions which directly correspond to
4833 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
4834 is needed for applications which have to configure certificates on a
4835 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
4837 For the RSA certificate situation is makes no difference, but
4838 for the DSA certificate situation this fixes the "no shared cipher"
4839 problem where the OpenSSL cipher selection procedure failed because the
4840 temporary keys were not overtaken from the context and the API provided
4841 no way to reconfigure them.
4842 The new functions now let applications reconfigure the stuff and they
4843 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
4844 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
4845 non-public-API function ssl_cert_instantiate() is used as a helper
4846 function and also to reduce code redundancy inside ssl_rsa.c.
4847 [Ralf S. Engelschall]
4849 *) Move s_server -dcert and -dkey options out of the undocumented feature
4850 area because they are useful for the DSA situation and should be
4851 recognized by the users.
4852 [Ralf S. Engelschall]
4854 *) Fix the cipher decision scheme for export ciphers: the export bits are
4855 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
4856 SSL_EXP_MASK. So, the original variable has to be used instead of the
4857 already masked variable.
4858 [Richard Levitte <levitte@stacken.kth.se>]
4860 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
4861 [Richard Levitte <levitte@stacken.kth.se>]
4863 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
4864 from `int' to `unsigned int' because it's a length and initialized by
4865 EVP_DigestFinal() which expects an `unsigned int *'.
4866 [Richard Levitte <levitte@stacken.kth.se>]
4868 *) Don't hard-code path to Perl interpreter on shebang line of Configure
4869 script. Instead use the usual Shell->Perl transition trick.
4870 [Ralf S. Engelschall]
4872 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
4873 (in addition to RSA certificates) to match the behaviour of `openssl dsa
4874 -noout -modulus' as it's already the case for `openssl rsa -noout
4875 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
4876 currently the public key is printed (a decision which was already done by
4877 `openssl dsa -modulus' in the past) which serves a similar purpose.
4878 Additionally the NO_RSA no longer completely removes the whole -modulus
4879 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
4881 [Ralf S. Engelschall]
4883 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
4884 BIO. See the source (crypto/evp/bio_ok.c) for more info.
4885 [Arne Ansper <arne@ats.cyber.ee>]
4887 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
4888 to be added. Now both 'req' and 'ca' can use new objects defined in the
4892 *) Add cool BIO that does syslog (or event log on NT).
4893 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
4895 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
4896 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
4897 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
4898 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
4901 *) Add preliminary config info for new extension code.
4904 *) Make RSA_NO_PADDING really use no padding.
4905 [Ulf Moeller <ulf@fitug.de>]
4907 *) Generate errors when private/public key check is done.
4910 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
4911 for some CRL extensions and new objects added.
4914 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
4915 key usage extension and fuller support for authority key id.
4918 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
4919 padding method for RSA, which is recommended for new applications in PKCS
4920 #1 v2.0 (RFC 2437, October 1998).
4921 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
4922 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
4923 against Bleichbacher's attack on RSA.
4924 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
4927 *) Updates to the new SSL compression code
4928 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
4930 *) Fix so that the version number in the master secret, when passed
4931 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
4932 (because the server will not accept higher), that the version number
4933 is 0x03,0x01, not 0x03,0x00
4934 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
4936 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
4937 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
4938 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
4941 *) Support for RAW extensions where an arbitrary extension can be
4942 created by including its DER encoding. See apps/openssl.cnf for
4946 *) Make sure latest Perl versions don't interpret some generated C array
4947 code as Perl array code in the crypto/err/err_genc.pl script.
4948 [Lars Weber <3weber@informatik.uni-hamburg.de>]
4950 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
4951 not many people have the assembler. Various Win32 compilation fixes and
4952 update to the INSTALL.W32 file with (hopefully) more accurate Win32
4956 *) Modify configure script 'Configure' to automatically create crypto/date.h
4957 file under Win32 and also build pem.h from pem.org. New script
4958 util/mkfiles.pl to create the MINFO file on environments that can't do a
4959 'make files': perl util/mkfiles.pl >MINFO should work.
4962 *) Major rework of DES function declarations, in the pursuit of correctness
4963 and purity. As a result, many evil casts evaporated, and some weirdness,
4964 too. You may find this causes warnings in your code. Zapping your evil
4965 casts will probably fix them. Mostly.
4968 *) Fix for a typo in asn1.h. Bug fix to object creation script
4969 obj_dat.pl. It considered a zero in an object definition to mean
4970 "end of object": none of the objects in objects.h have any zeros
4971 so it wasn't spotted.
4972 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
4974 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
4975 Masking (CBCM). In the absence of test vectors, the best I have been able
4976 to do is check that the decrypt undoes the encrypt, so far. Send me test
4977 vectors if you have them.
4980 *) Correct calculation of key length for export ciphers (too much space was
4981 allocated for null ciphers). This has not been tested!
4984 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
4985 message is now correct (it understands "crypto" and "ssl" on its
4986 command line). There is also now an "update" option. This will update
4987 the util/ssleay.num and util/libeay.num files with any new functions.
4989 perl util/mkdef.pl crypto ssl update
4990 it will update them.
4993 *) Overhauled the Perl interface (perl/*):
4994 - ported BN stuff to OpenSSL's different BN library
4995 - made the perl/ source tree CVS-aware
4996 - renamed the package from SSLeay to OpenSSL (the files still contain
4997 their history because I've copied them in the repository)
4998 - removed obsolete files (the test scripts will be replaced
4999 by better Test::Harness variants in the future)
5000 [Ralf S. Engelschall]
5002 *) First cut for a very conservative source tree cleanup:
5003 1. merge various obsolete readme texts into doc/ssleay.txt
5004 where we collect the old documents and readme texts.
5005 2. remove the first part of files where I'm already sure that we no
5006 longer need them because of three reasons: either they are just temporary
5007 files which were left by Eric or they are preserved original files where
5008 I've verified that the diff is also available in the CVS via "cvs diff
5009 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
5010 the crypto/md/ stuff).
5011 [Ralf S. Engelschall]
5013 *) More extension code. Incomplete support for subject and issuer alt
5014 name, issuer and authority key id. Change the i2v function parameters
5015 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
5016 what that's for :-) Fix to ASN1 macro which messed up
5017 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
5020 *) Preliminary support for ENUMERATED type. This is largely copied from the
5024 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
5025 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
5027 *) Make sure `make rehash' target really finds the `openssl' program.
5028 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
5030 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
5031 like to hear about it if this slows down other processors.
5034 *) Add CygWin32 platform information to Configure script.
5035 [Alan Batie <batie@aahz.jf.intel.com>]
5037 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
5038 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
5040 *) New program nseq to manipulate netscape certificate sequences
5043 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
5047 *) Fixes to BN code. Previously the default was to define BN_RECURSION
5048 but the BN code had some problems that would cause failures when
5049 doing certificate verification and some other functions.
5050 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
5052 *) Add ASN1 and PEM code to support netscape certificate sequences.
5055 *) Add ASN1 and PEM code to support netscape certificate sequences.
5058 *) Add several PKIX and private extended key usage OIDs.
5061 *) Modify the 'ca' program to handle the new extension code. Modify
5062 openssl.cnf for new extension format, add comments.
5065 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
5066 and add a sample to openssl.cnf so req -x509 now adds appropriate
5070 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
5071 error code, add initial support to X509_print() and x509 application.
5074 *) Takes a deep breath and start addding X509 V3 extension support code. Add
5075 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
5076 stuff is currently isolated and isn't even compiled yet.
5079 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
5080 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
5081 Removed the versions check from X509 routines when loading extensions:
5082 this allows certain broken certificates that don't set the version
5083 properly to be processed.
5086 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
5087 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
5088 can still be regenerated with "make depend".
5091 *) Spelling mistake in C version of CAST-128.
5092 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
5094 *) Changes to the error generation code. The perl script err-code.pl
5095 now reads in the old error codes and retains the old numbers, only
5096 adding new ones if necessary. It also only changes the .err files if new
5097 codes are added. The makefiles have been modified to only insert errors
5098 when needed (to avoid needlessly modifying header files). This is done
5099 by only inserting errors if the .err file is newer than the auto generated
5100 C file. To rebuild all the error codes from scratch (the old behaviour)
5101 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
5102 or delete all the .err files.
5105 *) CAST-128 was incorrectly implemented for short keys. The C version has
5106 been fixed, but is untested. The assembler versions are also fixed, but
5107 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
5108 to regenerate it if needed.
5109 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
5110 Hagino <itojun@kame.net>]
5112 *) File was opened incorrectly in randfile.c.
5113 [Ulf Möller <ulf@fitug.de>]
5115 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
5116 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
5117 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
5118 al: it's just almost always a UTCTime. Note this patch adds new error
5119 codes so do a "make errors" if there are problems.
5122 *) Correct Linux 1 recognition in config.
5123 [Ulf Möller <ulf@fitug.de>]
5125 *) Remove pointless MD5 hash when using DSA keys in ca.
5126 [Anonymous <nobody@replay.com>]
5128 *) Generate an error if given an empty string as a cert directory. Also
5129 generate an error if handed NULL (previously returned 0 to indicate an
5130 error, but didn't set one).
5131 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
5133 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
5136 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
5137 parameters. This was causing a warning which killed off the Win32 compile.
5140 *) Remove C++ style comments from crypto/bn/bn_local.h.
5141 [Neil Costigan <neil.costigan@celocom.com>]
5143 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
5144 based on a text string, looking up short and long names and finally
5145 "dot" format. The "dot" format stuff didn't work. Added new function
5146 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
5147 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
5148 OID is not part of the table.
5151 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
5152 X509_LOOKUP_by_alias().
5155 *) Sort openssl functions by name.
5158 *) Get the gendsa program working (hopefully) and add it to app list. Remove
5159 encryption from sample DSA keys (in case anyone is interested the password
5163 *) Make _all_ *_free functions accept a NULL pointer.
5164 [Frans Heymans <fheymans@isaserver.be>]
5166 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
5168 [Anonymous <nobody@replay.com>]
5170 *) s_server should send the CAfile as acceptable CAs, not its own cert.
5171 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
5173 *) Don't blow it for numeric -newkey arguments to apps/req.
5174 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
5176 *) Temp key "for export" tests were wrong in s3_srvr.c.
5177 [Anonymous <nobody@replay.com>]
5179 *) Add prototype for temp key callback functions
5180 SSL_CTX_set_tmp_{rsa,dh}_callback().
5183 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
5184 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
5187 *) X509_name_add_entry() freed the wrong thing after an error.
5188 [Arne Ansper <arne@ats.cyber.ee>]
5190 *) rsa_eay.c would attempt to free a NULL context.
5191 [Arne Ansper <arne@ats.cyber.ee>]
5193 *) BIO_s_socket() had a broken should_retry() on Windoze.
5194 [Arne Ansper <arne@ats.cyber.ee>]
5196 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
5197 [Arne Ansper <arne@ats.cyber.ee>]
5199 *) Make sure the already existing X509_STORE->depth variable is initialized
5200 in X509_STORE_new(), but document the fact that this variable is still
5201 unused in the certificate verification process.
5202 [Ralf S. Engelschall]
5204 *) Fix the various library and apps files to free up pkeys obtained from
5205 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
5208 *) Fix reference counting in X509_PUBKEY_get(). This makes
5209 demos/maurice/example2.c work, amongst others, probably.
5210 [Steve Henson and Ben Laurie]
5212 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
5213 `openssl' and second, the shortcut symlinks for the `openssl <command>'
5214 are no longer created. This way we have a single and consistent command
5215 line interface `openssl <command>', similar to `cvs <command>'.
5216 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
5218 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
5219 BIT STRING wrapper always have zero unused bits.
5222 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
5225 *) Make the top-level INSTALL documentation easier to understand.
5228 *) Makefiles updated to exit if an error occurs in a sub-directory
5229 make (including if user presses ^C) [Paul Sutton]
5231 *) Make Montgomery context stuff explicit in RSA data structure.
5234 *) Fix build order of pem and err to allow for generated pem.h.
5237 *) Fix renumbering bug in X509_NAME_delete_entry().
5240 *) Enhanced the err-ins.pl script so it makes the error library number
5241 global and can add a library name. This is needed for external ASN1 and
5242 other error libraries.
5245 *) Fixed sk_insert which never worked properly.
5248 *) Fix ASN1 macros so they can handle indefinite length construted
5249 EXPLICIT tags. Some non standard certificates use these: they can now
5253 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
5254 into a single doc/ssleay.txt bundle. This way the information is still
5255 preserved but no longer messes up this directory. Now it's new room for
5256 the new set of documenation files.
5257 [Ralf S. Engelschall]
5259 *) SETs were incorrectly DER encoded. This was a major pain, because they
5260 shared code with SEQUENCEs, which aren't coded the same. This means that
5261 almost everything to do with SETs or SEQUENCEs has either changed name or
5262 number of arguments.
5263 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
5265 *) Fix test data to work with the above.
5268 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
5269 was already fixed by Eric for 0.9.1 it seems.
5270 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
5272 *) Autodetect FreeBSD3.
5275 *) Fix various bugs in Configure. This affects the following platforms:
5279 unixware-2.0-pentium
5283 *) Eliminate generated files from CVS. Reorder tests to regenerate files
5284 before they are needed.
5287 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
5291 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
5293 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
5294 changed SSLeay to OpenSSL in version strings.
5295 [Ralf S. Engelschall]
5297 *) Some fixups to the top-level documents.
5300 *) Fixed the nasty bug where rsaref.h was not found under compile-time
5301 because the symlink to include/ was missing.
5302 [Ralf S. Engelschall]
5304 *) Incorporated the popular no-RSA/DSA-only patches
5305 which allow to compile a RSA-free SSLeay.
5306 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
5308 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
5309 when "ssleay" is still not found.
5310 [Ralf S. Engelschall]
5312 *) Added more platforms to Configure: Cray T3E, HPUX 11,
5313 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
5315 *) Updated the README file.
5316 [Ralf S. Engelschall]
5318 *) Added various .cvsignore files in the CVS repository subdirs
5319 to make a "cvs update" really silent.
5320 [Ralf S. Engelschall]
5322 *) Recompiled the error-definition header files and added
5323 missing symbols to the Win32 linker tables.
5324 [Ralf S. Engelschall]
5326 *) Cleaned up the top-level documents;
5327 o new files: CHANGES and LICENSE
5328 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
5329 o merged COPYRIGHT into LICENSE
5330 o removed obsolete TODO file
5331 o renamed MICROSOFT to INSTALL.W32
5332 [Ralf S. Engelschall]
5334 *) Removed dummy files from the 0.9.1b source tree:
5335 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
5336 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
5337 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
5338 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
5339 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
5340 [Ralf S. Engelschall]
5342 *) Added various platform portability fixes.
5345 *) The Genesis of the OpenSSL rpject:
5346 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
5347 Young and Tim J. Hudson created while they were working for C2Net until
5349 [The OpenSSL Project]
5352 Changes between 0.9.0b and 0.9.1b [not released]
5354 *) Updated a few CA certificates under certs/
5357 *) Changed some BIGNUM api stuff.
5360 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
5361 DGUX x86, Linux Alpha, etc.
5364 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
5365 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
5369 *) Add -strparse option to asn1pars program which parses nested
5371 [Dr Stephen Henson <shenson@bigfoot.com>]
5373 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
5376 *) DSA fix for "ca" program.
5379 *) Added "-genkey" option to "dsaparam" program.
5382 *) Added RIPE MD160 (rmd160) message digest.
5385 *) Added -a (all) option to "ssleay version" command.
5388 *) Added PLATFORM define which is the id given to Configure.
5391 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
5394 *) Extended the ASN.1 parser routines.
5397 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
5400 *) Added a BN_CTX to the BN library.
5403 *) Fixed the weak key values in DES library
5406 *) Changed API in EVP library for cipher aliases.
5409 *) Added support for RC2/64bit cipher.
5412 *) Converted the lhash library to the crypto/mem.c functions.
5415 *) Added more recognized ASN.1 object ids.
5418 *) Added more RSA padding checks for SSL/TLS.
5421 *) Added BIO proxy/filter functionality.
5424 *) Added extra_certs to SSL_CTX which can be used
5425 send extra CA certificates to the client in the CA cert chain sending
5426 process. It can be configured with SSL_CTX_add_extra_chain_cert().
5429 *) Now Fortezza is denied in the authentication phase because
5430 this is key exchange mechanism is not supported by SSLeay at all.
5433 *) Additional PKCS1 checks.
5436 *) Support the string "TLSv1" for all TLS v1 ciphers.
5439 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
5440 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
5443 *) Fixed a few memory leaks.
5446 *) Fixed various code and comment typos.
5449 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
5450 bytes sent in the client random.
5451 [Edward Bishop <ebishop@spyglass.com>]