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 +) Modify the behaviour of EVP cipher functions in similar way to digests
16 to retain compatibility with existing code.
19 +) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
20 compatibility with existing code. In particular the 'ctx' parameter is
21 not assumed to be valid before the call to EVP_DigestInit() and it is tidied
22 up after a call to EVP_DigestFinal(). A new function EVP_DigestFinal_ex()
23 but does not free up the ctx. Also change function EVP_MD_CTX_copy() to
24 assume the destination is uninitialized: EVP_MD_CTX_copy_ex() do assumes
25 the destiation is valid. Also modify all the OpenSSL digest calls to call
26 EVP_DigestInit_ex(), EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
29 +) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
30 so that complete 'Handshake' protocol structures are kept in memory
31 instead of overwriting 'msg_type' and 'length' with 'body' data.
34 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
38 +) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
39 [Massimo Santin via Richard Levitte]
41 +) Major restructuring to the underlying ENGINE code. This includes
42 reduction of linker bloat, separation of pure "ENGINE" manipulation
43 (initialisation, etc) from functionality dealing with implementations
44 of specific crypto iterfaces. This change also introduces integrated
45 support for symmetric ciphers and digest implementations - so ENGINEs
46 can now accelerate these by providing EVP_CIPHER and EVP_MD
47 implementations of their own. This is detailed in crypto/engine/README
48 as it couldn't be adequately described here. However, there are a few
49 API changes worth noting - some RSA, DSA, DH, and RAND functions that
50 were changed in the original introduction of ENGINE code have now
51 reverted back - the hooking from this code to ENGINE is now a good
52 deal more passive and at run-time, operations deal directly with
53 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
54 dereferencing through an ENGINE pointer any more. Also, the ENGINE
55 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
56 they were not being used by the framework as there is no concept of a
57 BIGNUM_METHOD and they could not be generalised to the new
58 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
59 ENGINE_cpy() has been removed as it cannot be consistently defined in
63 +) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
66 +) Change mkdef.pl to sort symbols that get the same entry number,
67 and make sure the automatically generated functions ERR_load_*
68 become part of libeay.num as well.
71 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
72 client receives HelloRequest while in a handshake.
73 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
75 +) New function SSL_renegotiate_pending(). This returns true once
76 renegotiation has been requested (either SSL_renegotiate() call
77 or HelloRequest/ClientHello receveived from the peer) and becomes
78 false once a handshake has been completed.
79 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
80 sends a HelloRequest, but does not ensure that a handshake takes
81 place. SSL_renegotiate_pending() is useful for checking if the
82 client has followed the request.)
85 +) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
86 By default, clients may request session resumption even during
87 renegotiation (if session ID contexts permit); with this option,
88 session resumption is possible only in the first handshake.
91 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
92 should end in 'break', not 'goto end' which circuments various
93 cleanups done in state SSL_ST_OK. But session related stuff
94 must be disabled for SSL_ST_OK in the case that we just sent a
97 Also avoid some overhead by not calling ssl_init_wbio_buffer()
98 before just sending a HelloRequest.
99 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
101 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
102 reveal whether illegal block cipher padding was found or a MAC
103 verification error occured. (Neither SSLerr() codes nor alerts
104 are directly visible to potential attackers, but the information
105 may leak via logfiles.)
107 Similar changes are not required for the SSL 2.0 implementation
108 because the number of padding bytes is sent in clear for SSL 2.0,
109 and the extra bytes are just ignored. However ssl/s2_pkt.c
110 failed to verify that the purported number of padding bytes is in
114 +) Add some demos for certificate and certificate request creation.
117 +) Make maximum certificate chain size accepted from the peer application
118 settable (SSL*_get/set_max_cert_list()), as proposed by
119 "Douglas E. Engert" <deengert@anl.gov>.
122 +) Add support for shared libraries for Unixware-7 and support including
123 shared libraries for OpenUNIX-8 (Boyd Lynn Gerber <gerberb@zenez.com>).
126 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
127 'wristwatch attack' using huge encoding parameters (cf.
128 James H. Manger's CRYPTO 2001 paper). Note that the
129 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
130 encoding parameters and hence was not vulnerable.
133 +) Add a "destroy" handler to ENGINEs that allows structural cleanup to
134 be done prior to destruction. Use this to unload error strings from
135 ENGINEs that load their own error strings. NB: This adds two new API
136 functions to "get" and "set" this destroy handler in an ENGINE.
139 +) Alter all existing ENGINE implementations (except "openssl" and
140 "openbsd") to dynamically instantiate their own error strings. This
141 makes them more flexible to be built both as statically-linked ENGINEs
142 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
143 Also, add stub code to each that makes building them as self-contained
144 shared-libraries easier (see README.ENGINE).
147 +) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
148 implementations into applications that are completely implemented in
149 self-contained shared-libraries. The "dynamic" ENGINE exposes control
150 commands that can be used to configure what shared-library to load and
151 to control aspects of the way it is handled. Also, made an update to
152 the README.ENGINE file that brings its information up-to-date and
153 provides some information and instructions on the "dynamic" ENGINE
154 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
158 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
160 *) Make it possible to unload ranges of ERR strings with a new
161 "ERR_unload_strings" function.
164 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
165 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
166 followed by modular reduction.
167 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
169 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
170 equivalent based on BN_pseudo_rand() instead of BN_rand().
173 +) Add a copy() function to EVP_MD.
176 +) Make EVP_MD routines take a context pointer instead of just the
177 md_data void pointer.
180 +) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
181 that the digest can only process a single chunk of data
182 (typically because it is provided by a piece of
183 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
184 is only going to provide a single chunk of data, and hence the
185 framework needn't accumulate the data for oneshot drivers.
188 +) As with "ERR", make it possible to replace the underlying "ex_data"
189 functions. This change also alters the storage and management of global
190 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
191 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
192 index counters. The API functions that use this state have been changed
193 to take a "class_index" rather than pointers to the class's local STACK
194 and counter, and there is now an API function to dynamically create new
195 classes. This centralisation allows us to (a) plug a lot of the
196 thread-safety problems that existed, and (b) makes it possible to clean
197 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
198 such data would previously have always leaked in application code and
199 workarounds were in place to make the memory debugging turn a blind eye
200 to it. Application code that doesn't use this new function will still
201 leak as before, but their memory debugging output will announce it now
202 rather than letting it slide.
204 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
205 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
206 has a return value to indicate success or failure.
209 +) Make it possible to replace the underlying "ERR" functions such that the
210 global state (2 LHASH tables and 2 locks) is only used by the "default"
211 implementation. This change also adds two functions to "get" and "set"
212 the implementation prior to it being automatically set the first time
213 any other ERR function takes place. Ie. an application can call "get",
214 pass the return value to a module it has just loaded, and that module
215 can call its own "set" function using that value. This means the
216 module's "ERR" operations will use (and modify) the error state in the
217 application and not in its own statically linked copy of OpenSSL code.
220 +) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
221 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
222 the operation, and provides a more encapsulated way for external code
223 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
224 to use these functions rather than manually incrementing the counts.
226 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
229 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
230 This function was broken, as the check for a new client hello message
231 to handle SGC did not allow these large messages.
232 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
235 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
238 +) Add EVP test program.
241 +) Add symmetric cipher support to ENGINE. Expect the API to change!
244 +) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
245 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
246 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
247 These allow a CRL to be built without having to access X509_CRL fields
248 directly. Modify 'ca' application to use new functions.
251 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
252 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
255 *) Rework the configuration and shared library support for Tru64 Unix.
256 The configuration part makes use of modern compiler features and
257 still retains old compiler behavior for those that run older versions
258 of the OS. The shared library support part includes a variant that
259 uses the RPATH feature, and is available through the special
260 configuration target "alpha-cc-rpath", which will never be selected
262 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
264 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
265 with the same message size as in ssl3_get_certificate_request().
266 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
267 messages might inadvertently be reject as too long.
268 [Petr Lampa <lampa@fee.vutbr.cz>]
270 +) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
271 bug workarounds. Rollback attack detection is a security feature.
272 The problem will only arise on OpenSSL servers when TLSv1 is not
273 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
274 Software authors not wanting to support TLSv1 will have special reasons
275 for their choice and can explicitly enable this option.
276 [Bodo Moeller, Lutz Jaenicke]
278 +) Rationalise EVP so it can be extended: don't include a union of
279 cipher/digest structures, add init/cleanup functions. This also reduces
280 the number of header dependencies.
283 +) Make DES key schedule conform to the usual scheme, as well as
284 correcting its structure. This means that calls to DES functions
285 now have to pass a pointer to a des_key_schedule instead of a
286 plain des_key_schedule (which was actually always a pointer
290 +) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
293 *) Modified SSL library such that the verify_callback that has been set
294 specificly for an SSL object with SSL_set_verify() is actually being
295 used. Before the change, a verify_callback set with this function was
296 ignored and the verify_callback() set in the SSL_CTX at the time of
297 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
298 to allow the necessary settings.
301 +) Initial reduction of linker bloat: the use of some functions, such as
302 PEM causes large amounts of unused functions to be linked in due to
303 poor organisation. For example pem_all.c contains every PEM function
304 which has a knock on effect of linking in large amounts of (unused)
305 ASN1 code. Grouping together similar functions and splitting unrelated
306 functions prevents this.
309 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
310 explicitly to NULL, as at least on Solaris 8 this seems not always to be
311 done automatically (in contradiction to the requirements of the C
312 standard). This made problems when used from OpenSSH.
315 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
316 dh->length and always used
318 BN_rand_range(priv_key, dh->p).
320 BN_rand_range() is not necessary for Diffie-Hellman, and this
321 specific range makes Diffie-Hellman unnecessarily inefficient if
322 dh->length (recommended exponent length) is much smaller than the
323 length of dh->p. We could use BN_rand_range() if the order of
324 the subgroup was stored in the DH structure, but we only have
329 BN_rand(priv_key, l, ...)
331 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
337 RSA_eay_public_encrypt
338 RSA_eay_private_decrypt
339 RSA_eay_private_encrypt (signing)
340 RSA_eay_public_decrypt (signature verification)
342 (default implementations for RSA_public_encrypt,
343 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
344 always reject numbers >= n.
347 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
348 to synchronize access to 'locking_thread'. This is necessary on
349 systems where access to 'locking_thread' (an 'unsigned long'
350 variable) is not atomic.
353 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
354 *before* setting the 'crypto_lock_rand' flag. The previous code had
355 a race condition if 0 is a valid thread ID.
356 [Travis Vitek <vitek@roguewave.com>]
358 +) Cleanup of EVP macros.
361 +) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
365 +) Add initial OCSP responder support to ocsp application. The
366 revocation information is handled using the text based index
367 use by the ca application. The responder can either handle
368 requests generated internally, supplied in files (for example
369 via a CGI script) or using an internal minimal server.
372 +) Add configuration choices to get zlib compression for TLS.
375 +) Changes to Kerberos SSL for RFC 2712 compliance:
376 1. Implemented real KerberosWrapper, instead of just using
377 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
378 2. Implemented optional authenticator field of KerberosWrapper.
380 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
381 and authenticator structs; see crypto/krb5/.
383 Generalized Kerberos calls to support multiple Kerberos libraries.
384 [Vern Staats <staatsvr@asc.hpc.mil>,
385 Jeffrey Altman <jaltman@columbia.edu>
388 +) Cause 'openssl speed' to use fully hard-coded DSA keys as it
389 already does with RSA. testdsa.h now has 'priv_key/pub_key'
390 values for each of the key sizes rather than having just
391 parameters (and 'speed' generating keys each time).
394 -) OpenSSL 0.9.6b released [9 July 2001]
396 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
397 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
398 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
399 PRNG state recovery was possible based on the output of
400 one PRNG request appropriately sized to gain knowledge on
401 'md' followed by enough consecutive 1-byte PRNG requests
402 to traverse all of 'state'.
404 1. When updating 'md_local' (the current thread's copy of 'md')
405 during PRNG output generation, hash all of the previous
406 'md_local' value, not just the half used for PRNG output.
408 2. Make the number of bytes from 'state' included into the hash
409 independent from the number of PRNG bytes requested.
411 The first measure alone would be sufficient to avoid
412 Markku-Juhani's attack. (Actually it had never occurred
413 to me that the half of 'md_local' used for chaining was the
414 half from which PRNG output bytes were taken -- I had always
415 assumed that the secret half would be used.) The second
416 measure makes sure that additional data from 'state' is never
417 mixed into 'md_local' in small portions; this heuristically
418 further strengthens the PRNG.
421 +) Speed up EVP routines.
424 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
425 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
426 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
427 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
429 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
430 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
431 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
434 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
436 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
439 *) Fix crypto/bn/asm/mips3.s.
442 *) When only the key is given to "enc", the IV is undefined. Print out
443 an error message in this case.
446 +) Added the OS2-EMX target.
447 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
449 +) Rewrite apps to use NCONF routines instead of the old CONF. New functions
450 to support NCONF routines in extension code. New function CONF_set_nconf()
451 to allow functions which take an NCONF to also handle the old LHASH
452 structure: this means that the old CONF compatible routines can be
453 retained (in particular wrt extensions) without having to duplicate the
454 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
457 *) Handle special case when X509_NAME is empty in X509 printing routines.
460 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
461 positive and less than q.
464 +) Enhance the general user interface with mechanisms for inner control
465 and with possibilities to have yes/no kind of prompts.
468 +) Change all calls to low level digest routines in the library and
469 applications to use EVP. Add missing calls to HMAC_cleanup() and
470 don't assume HMAC_CTX can be copied using memcpy().
471 [Verdon Walker <VWalker@novell.com>, Steve Henson]
473 +) Add the possibility to control engines through control names but with
474 arbitrary arguments instead of just a string.
475 Change the key loaders to take a UI_METHOD instead of a callback
476 function pointer. NOTE: this breaks binary compatibility with earlier
477 versions of OpenSSL [engine].
478 Adapt the nCipher code for these new conditions and add a card insertion
482 +) Enhance the general user interface with mechanisms to better support
483 dialog box interfaces, application-defined prompts, the possibility
484 to use defaults (for example default passwords from somewhere else)
485 and interrupts/cancellations.
488 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
489 used: it isn't thread safe and the add_lock_callback should handle
491 [Paul Rose <Paul.Rose@bridge.com>]
493 *) Verify that incoming data obeys the block size in
494 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
497 +) Tidy up PKCS#12 attribute handling. Add support for the CSP name
498 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
502 [Ulf Möller, Bodo Möller]
504 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
505 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
506 when fixing the server behaviour for backwards-compatible 'client
507 hello' messages. (Note that the attack is impractical against
508 SSL 3.0 and TLS 1.0 anyway because length and version checking
509 means that the probability of guessing a valid ciphertext is
510 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
513 Before 0.9.5, the countermeasure (hide the error by generating a
514 random 'decryption result') did not work properly because
515 ERR_clear_error() was missing, meaning that SSL_get_error() would
516 detect the supposedly ignored error.
518 Both problems are now fixed.
521 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
522 (previously it was 1024).
525 +) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
526 tidy up some unnecessarily weird code in 'sk_new()').
527 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
529 +) Change the key loading routines for ENGINEs to use the same kind
530 callback (pem_password_cb) as all other routines that need this
534 *) Fix for compatibility mode trust settings: ignore trust settings
535 unless some valid trust or reject settings are present.
538 *) Fix for blowfish EVP: its a variable length cipher.
541 +) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
542 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
543 than this minimum value is recommended.
546 +) New random seeder for OpenVMS, using the system process statistics
547 that are easily reachable.
550 +) Windows apparently can't transparently handle global
551 variables defined in DLLs. Initialisations such as:
553 const ASN1_ITEM *it = &ASN1_INTEGER_it;
555 wont compile. This is used by the any applications that need to
556 declare their own ASN1 modules. This was fixed by adding the option
557 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
558 needed for static libraries under Win32.
561 +) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
562 setting of purpose and trust fields. New X509_STORE trust and
563 purpose functions and tidy up setting in other SSL functions.
566 +) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
567 structure. These are inherited by X509_STORE_CTX when it is
568 initialised. This allows various defaults to be set in the
569 X509_STORE structure (such as flags for CRL checking and custom
570 purpose or trust settings) for functions which only use X509_STORE_CTX
571 internally such as S/MIME.
573 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
574 trust settings if they are not set in X509_STORE. This allows X509_STORE
575 purposes and trust (in S/MIME for example) to override any set by default.
577 Add command line options for CRL checking to smime, s_client and s_server
581 +) Initial CRL based revocation checking. If the CRL checking flag(s)
582 are set then the CRL is looked up in the X509_STORE structure and
583 its validity and signature checked, then if the certificate is found
584 in the CRL the verify fails with a revoked error.
586 Various new CRL related callbacks added to X509_STORE_CTX structure.
588 Command line options added to 'verify' application to support this.
590 This needs some additional work, such as being able to handle multiple
591 CRLs with different times, extension based lookup (rather than just
592 by subject name) and ultimately more complete V2 CRL extension
596 +) Add a general user interface API (crypto/ui/). This is designed
597 to replace things like des_read_password and friends (backward
598 compatibility functions using this new API are provided).
599 The purpose is to remove prompting functions from the DES code
600 section as well as provide for prompting through dialog boxes in
601 a window system and the like.
604 *) Fix various bugs related to DSA S/MIME verification. Handle missing
605 parameters in DSA public key structures and return an error in the
606 DSA routines if parameters are absent.
609 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
610 in the current directory if neither $RANDFILE nor $HOME was set.
611 RAND_file_name() in 0.9.6a returned NULL in this case. This has
612 caused some confusion to Windows users who haven't defined $HOME.
613 Thus RAND_file_name() is changed again: e_os.h can define a
614 DEFAULT_HOME, which will be used if $HOME is not set.
615 For Windows, we use "C:"; on other platforms, we still require
616 environment variables.
618 +) Add "ex_data" support to ENGINE so implementations can add state at a
619 per-structure level rather than having to store it globally.
622 +) Make it possible for ENGINE structures to be copied when retrieved by
623 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
624 This causes the "original" ENGINE structure to act like a template,
625 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
626 operational state can be localised to each ENGINE structure, despite the
627 fact they all share the same "methods". New ENGINE structures returned in
628 this case have no functional references and the return value is the single
629 structural reference. This matches the single structural reference returned
630 by ENGINE_by_id() normally, when it is incremented on the pre-existing
634 +) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
635 needs to match any other type at all we need to manually clear the
639 +) Changes to the "openssl engine" utility to include;
640 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
641 about an ENGINE's available control commands.
642 - executing control commands from command line arguments using the
643 '-pre' and '-post' switches. '-post' is only used if '-t' is
644 specified and the ENGINE is successfully initialised. The syntax for
645 the individual commands are colon-separated, for example;
646 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
649 +) New dynamic control command support for ENGINEs. ENGINEs can now
650 declare their own commands (numbers), names (strings), descriptions,
651 and input types for run-time discovery by calling applications. A
652 subset of these commands are implicitly classed as "executable"
653 depending on their input type, and only these can be invoked through
654 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
655 can be based on user input, config files, etc). The distinction is
656 that "executable" commands cannot return anything other than a boolean
657 result and can only support numeric or string input, whereas some
658 discoverable commands may only be for direct use through
659 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
660 pointers, or other custom uses. The "executable" commands are to
661 support parameterisations of ENGINE behaviour that can be
662 unambiguously defined by ENGINEs and used consistently across any
663 OpenSSL-based application. Commands have been added to all the
664 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
665 control over shared-library paths without source code alterations.
668 +) Changed all ENGINE implementations to dynamically allocate their
669 ENGINEs rather than declaring them statically. Apart from this being
670 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
671 this also allows the implementations to compile without using the
672 internal engine_int.h header.
675 +) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
676 'const' value. Any code that should be able to modify a RAND_METHOD
677 should already have non-const pointers to it (ie. they should only
678 modify their own ones).
681 +) Made a variety of little tweaks to the ENGINE code.
682 - "atalla" and "ubsec" string definitions were moved from header files
683 to C code. "nuron" string definitions were placed in variables
684 rather than hard-coded - allowing parameterisation of these values
685 later on via ctrl() commands.
686 - Removed unused "#if 0"'d code.
687 - Fixed engine list iteration code so it uses ENGINE_free() to release
688 structural references.
689 - Constified the RAND_METHOD element of ENGINE structures.
690 - Constified various get/set functions as appropriate and added
691 missing functions (including a catch-all ENGINE_cpy that duplicates
692 all ENGINE values onto a new ENGINE except reference counts/state).
693 - Removed NULL parameter checks in get/set functions. Setting a method
694 or function to NULL is a way of cancelling out a previously set
695 value. Passing a NULL ENGINE parameter is just plain stupid anyway
696 and doesn't justify the extra error symbols and code.
697 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
698 flags from engine_int.h to engine.h.
699 - Changed prototypes for ENGINE handler functions (init(), finish(),
700 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
703 *) Move 'if (!initialized) RAND_poll()' into regions protected by
704 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
705 having multiple threads call RAND_poll() concurrently.
708 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
709 combination of a flag and a thread ID variable.
710 Otherwise while one thread is in ssleay_rand_bytes (which sets the
711 flag), *other* threads can enter ssleay_add_bytes without obeying
712 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
713 that they do not hold after the first thread unsets add_do_not_lock).
716 +) Implement binary inversion algorithm for BN_mod_inverse in addition
717 to the algorithm using long division. The binary algorithm can be
718 used only if the modulus is odd. On 32-bit systems, it is faster
719 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
720 roughly 5-15% for 256-bit moduli), so we use it only for moduli
721 up to 450 bits. In 64-bit environments, the binary algorithm
722 appears to be advantageous for much longer moduli; here we use it
723 for moduli up to 2048 bits.
726 *) Change bctest again: '-x' expressions are not available in all
730 -) OpenSSL 0.9.6a released [5 Apr 2001]
732 *) Fix a couple of memory leaks in PKCS7_dataDecode()
733 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
735 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
736 the default extension for executables, if any. Also, make the perl
737 scripts that use symlink() to test if it really exists and use "cp"
738 if it doesn't. All this made OpenSSL compilable and installable in
742 +) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
743 could not support the combine flag in choice fields.
746 -) Fix for asn1_GetSequence() for indefinite length constructed data.
747 If SEQUENCE is length is indefinite just set c->slen to the total
748 amount of data available.
749 [Steve Henson, reported by shige@FreeBSD.org]
750 [This change does not apply to 0.9.7.]
752 *) Change bctest to avoid here-documents inside command substitution
753 (workaround for FreeBSD /bin/sh bug).
754 For compatibility with Ultrix, avoid shell functions (introduced
755 in the bctest version that searches along $PATH).
758 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
759 with des_encrypt() defined on some operating systems, like Solaris
763 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
764 On the Importance of Eliminating Errors in Cryptographic
765 Computations, J. Cryptology 14 (2001) 2, 101-119,
766 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
769 *) MIPS assembler BIGNUM division bug fix.
772 *) Disabled incorrect Alpha assembler code.
775 -) Fix PKCS#7 decode routines so they correctly update the length
776 after reading an EOC for the EXPLICIT tag.
778 [This change does not apply to 0.9.7.]
780 *) Fix bug in PKCS#12 key generation routines. This was triggered
781 if a 3DES key was generated with a 0 initial byte. Include
782 PKCS12_BROKEN_KEYGEN compilation option to retain the old
783 (but broken) behaviour.
786 *) Enhance bctest to search for a working bc along $PATH and print
788 [Tim Rice <tim@multitalents.net> via Richard Levitte]
790 +) Add a 'copy_extensions' option to the 'ca' utility. This copies
791 extensions from a certificate request to the certificate.
794 +) Allow multiple 'certopt' and 'nameopt' options to be separated
795 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
796 file: this allows the display of the certificate about to be
797 signed to be customised, to allow certain fields to be included
798 or excluded and extension details. The old system didn't display
799 multicharacter strings properly, omitted fields not in the policy
800 and couldn't display additional details such as extensions.
803 *) Fix memory leaks in err.c: free err_data string if necessary;
804 don't write to the wrong index in ERR_set_error_data.
807 +) Function EC_POINTs_mul for simultaneous scalar multiplication
808 of an arbitrary number of elliptic curve points, optionally
809 including the generator defined for the EC_GROUP.
810 EC_POINT_mul is a simple wrapper function for the typical case
811 that the point list has just one item (besides the optional
815 +) First EC_METHODs for curves over GF(p):
817 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
818 operations and provides various method functions that can also
819 operate with faster implementations of modular arithmetic.
821 EC_GFp_mont_method() reuses most functions that are part of
822 EC_GFp_simple_method, but uses Montgomery arithmetic.
824 [Bodo Moeller; point addition and point doubling
825 implementation directly derived from source code provided by
826 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
828 +) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
831 Curves are EC_GROUP objects (with an optional group generator)
832 based on EC_METHODs that are built into the library.
834 Points are EC_POINT objects based on EC_GROUP objects.
836 Most of the framework would be able to handle curves over arbitrary
837 finite fields, but as there are no obvious types for fields other
838 than GF(p), some functions are limited to that for now.
841 +) Add the -HTTP option to s_server. It is similar to -WWW, but requires
842 that the file contains a complete HTTP response.
845 +) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
846 change the def and num file printf format specifier from "%-40sXXX"
847 to "%-39s XXX". The latter will always guarantee a space after the
848 field while the former will cause them to run together if the field
849 is 40 of more characters long.
852 +) Constify the cipher and digest 'method' functions and structures
853 and modify related functions to take constant EVP_MD and EVP_CIPHER
857 *) Implement ssl23_peek (analogous to ssl23_read), which previously
861 *) Replace rdtsc with _emit statements for VC++ version 5.
862 [Jeremy Cooper <jeremy@baymoo.org>]
864 +) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
865 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
868 +) Modify EVP_Digest*() routines so they now return values. Although the
869 internal software routines can never fail additional hardware versions
873 +) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
875 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
876 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
879 ERR_R_NESTED_ASN1_ERROR
881 ERR_R_MISSING_ASN1_EOS
882 were 4 .. 9, conflicting with
883 ERR_LIB_RSA (= ERR_R_RSA_LIB)
885 ERR_LIB_PEM (= ERR_R_PEM_LIB).
886 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
888 Add new error code 'ERR_R_INTERNAL_ERROR'.
891 +) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
895 +) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
896 sets the subject name for a new request or supersedes the
897 subject name in a given request. Formats that can be parsed are
898 'CN=Some Name, OU=myOU, C=IT'
900 'CN=Some Name/OU=myOU/C=IT'.
902 Add options '-batch' and '-verbose' to 'openssl req'.
903 [Massimiliano Pala <madwolf@hackmasters.net>]
905 *) Make it possible to reuse SSLv2 sessions.
908 +) Introduce the possibility to access global variables through
909 functions on platform were that's the best way to handle exporting
910 global variables in shared libraries. To enable this functionality,
911 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
912 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
913 is normally done by Configure or something similar).
915 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
916 in the source file (foo.c) like this:
918 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
919 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
921 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
922 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
924 OPENSSL_DECLARE_GLOBAL(int,foo);
925 #define foo OPENSSL_GLOBAL_REF(foo)
926 OPENSSL_DECLARE_GLOBAL(double,bar);
927 #define bar OPENSSL_GLOBAL_REF(bar)
929 The #defines are very important, and therefore so is including the
930 header file everywhere where the defined globals are used.
932 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
933 of ASN.1 items, but that structure is a bit different.
935 The largest change is in util/mkdef.pl which has been enhanced with
936 better and easier to understand logic to choose which symbols should
937 go into the Windows .def files as well as a number of fixes and code
938 cleanup (among others, algorithm keywords are now sorted
939 lexicographically to avoid constant rewrites).
942 *) In copy_email() check for >= 0 as a return value for
943 X509_NAME_get_index_by_NID() since 0 is a valid index.
944 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
946 +) In BN_div() keep a copy of the sign of 'num' before writing the
947 result to 'rm' because if rm==num the value will be overwritten
948 and produce the wrong result if 'num' is negative: this caused
949 problems with BN_mod() and BN_nnmod().
952 +) Function OCSP_request_verify(). This checks the signature on an
953 OCSP request and verifies the signer certificate. The signer
954 certificate is just checked for a generic purpose and OCSP request
958 +) Add OCSP_check_validity() function to check the validity of OCSP
959 responses. OCSP responses are prepared in real time and may only
960 be a few seconds old. Simply checking that the current time lies
961 between thisUpdate and nextUpdate max reject otherwise valid responses
962 caused by either OCSP responder or client clock inaccuracy. Instead
963 we allow thisUpdate and nextUpdate to fall within a certain period of
964 the current time. The age of the response can also optionally be
965 checked. Two new options -validity_period and -status_age added to
969 +) If signature or public key algorithm is unrecognized print out its
970 OID rather that just UNKNOWN.
973 *) Avoid coredump with unsupported or invalid public keys by checking if
974 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
975 PKCS7_verify() fails with non detached data.
978 +) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
979 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
980 ID to be generated from the issuer certificate alone which can then be
981 passed to OCSP_id_issuer_cmp().
984 +) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
985 ASN1 modules to export functions returning ASN1_ITEM pointers
986 instead of the ASN1_ITEM structures themselves. This adds several
987 new macros which allow the underlying ASN1 function/structure to
988 be accessed transparently. As a result code should not use ASN1_ITEM
989 references directly (such as &X509_it) but instead use the relevant
990 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
991 use of the new ASN1 code on platforms where exporting structures
992 is problematical (for example in shared libraries) but exporting
993 functions returning pointers to structures is not.
996 +) Add support for overriding the generation of SSL/TLS session IDs.
997 These callbacks can be registered either in an SSL_CTX or per SSL.
998 The purpose of this is to allow applications to control, if they wish,
999 the arbitrary values chosen for use as session IDs, particularly as it
1000 can be useful for session caching in multiple-server environments. A
1001 command-line switch for testing this (and any client code that wishes
1002 to use such a feature) has been added to "s_server".
1003 [Geoff Thorpe, Lutz Jaenicke]
1005 +) Modify mkdef.pl to recognise and parse preprocessor conditionals
1006 of the form '#if defined(...) || defined(...) || ...' and
1007 '#if !defined(...) && !defined(...) && ...'. This also avoids
1008 the growing number of special cases it was previously handling.
1011 *) Don't use getenv in library functions when run as setuid/setgid.
1012 New function OPENSSL_issetugid().
1015 +) Make all configuration macros available for application by making
1016 sure they are available in opensslconf.h, by giving them names starting
1017 with "OPENSSL_" to avoid conflicts with other packages and by making
1018 sure e_os2.h will cover all platform-specific cases together with
1020 Additionally, it is now possible to define configuration/platform-
1021 specific names (called "system identities"). In the C code, these
1022 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
1023 macro with the name beginning with "OPENSSL_SYS_", which is determined
1024 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
1028 +) New option -set_serial to 'req' and 'x509' this allows the serial
1029 number to use to be specified on the command line. Previously self
1030 signed certificates were hard coded with serial number 0 and the
1031 CA options of 'x509' had to use a serial number in a file which was
1035 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
1036 due to incorrect handling of multi-threading:
1038 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
1040 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
1042 3. Count how many times MemCheck_off() has been called so that
1043 nested use can be treated correctly. This also avoids
1044 inband-signalling in the previous code (which relied on the
1045 assumption that thread ID 0 is impossible).
1048 +) New options to 'ca' utility to support V2 CRL entry extensions.
1049 Currently CRL reason, invalidity date and hold instruction are
1050 supported. Add new CRL extensions to V3 code and some new objects.
1053 *) Add "-rand" option also to s_client and s_server.
1056 +) New function EVP_CIPHER_CTX_set_padding() this is used to
1057 disable standard block padding (aka PKCS#5 padding) in the EVP
1058 API, which was previously mandatory. This means that the data is
1059 not padded in any way and so the total length much be a multiple
1060 of the block size, otherwise an error occurs.
1063 +) Initial (incomplete) OCSP SSL support.
1066 *) Fix CPU detection on Irix 6.x.
1067 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
1068 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
1070 +) New function OCSP_parse_url(). This splits up a URL into its host,
1071 port and path components: primarily to parse OCSP URLs. New -url
1072 option to ocsp utility.
1075 +) New nonce behavior. The return value of OCSP_check_nonce() now
1076 reflects the various checks performed. Applications can decide
1077 whether to tolerate certain situations such as an absent nonce
1078 in a response when one was present in a request: the ocsp application
1079 just prints out a warning. New function OCSP_add1_basic_nonce()
1080 this is to allow responders to include a nonce in a response even if
1081 the request is nonce-less.
1084 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
1087 [This change does not apply to 0.9.7.]
1089 *) Use the cached encoding of an X509_NAME structure rather than
1090 copying it. This is apparently the reason for the libsafe "errors"
1091 but the code is actually correct.
1094 +) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
1095 skipped when using openssl x509 multiple times on a single input file,
1096 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
1099 +) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
1100 set string type: to handle setting ASN1_TIME structures. Fix ca
1101 utility to correctly initialize revocation date of CRLs.
1104 +) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
1105 the clients preferred ciphersuites and rather use its own preferences.
1106 Should help to work around M$ SGC (Server Gated Cryptography) bug in
1107 Internet Explorer by ensuring unchanged hash method during stepup.
1108 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
1111 +) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
1112 to aes and add a new 'exist' option to print out symbols that don't
1116 +) Additional options to ocsp utility to allow flags to be set and
1117 additional certificates supplied.
1120 +) Add the option -VAfile to 'openssl ocsp', so the user can give the
1121 OCSP client a number of certificate to only verify the response
1125 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
1126 Bleichenbacher's DSA attack.
1127 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
1128 to be set and top=0 forces the highest bit to be set; top=-1 is new
1129 and leaves the highest bit random.
1130 [Ulf Moeller, Bodo Moeller]
1132 +) Update Rijndael code to version 3.0 and change EVP AES ciphers to
1133 handle the new API. Currently only ECB, CBC modes supported. Add new
1134 AES OIDs. Add TLS AES ciphersuites as described in the "AES Ciphersuites
1135 for TLS" draft-ietf-tls-ciphersuite-03.txt.
1136 [Ben Laurie, Steve Henson]
1138 *) In the NCONF_...-based implementations for CONF_... queries
1139 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
1140 a temporary CONF structure with the data component set to NULL
1141 (which gives segmentation faults in lh_retrieve).
1142 Instead, use NULL for the CONF pointer in CONF_get_string and
1143 CONF_get_number (which may use environment variables) and directly
1144 return NULL from CONF_get_section.
1147 *) Fix potential buffer overrun for EBCDIC.
1150 +) New function OCSP_copy_nonce() to copy nonce value (if present) from
1151 request to response.
1154 +) Functions for OCSP responders. OCSP_request_onereq_count(),
1155 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
1156 extract information from a certificate request. OCSP_response_create()
1157 creates a response and optionally adds a basic response structure.
1158 OCSP_basic_add1_status() adds a complete single response to a basic
1159 response and returns the OCSP_SINGLERESP structure just added (to allow
1160 extensions to be included for example). OCSP_basic_add1_cert() adds a
1161 certificate to a basic response and OCSP_basic_sign() signs a basic
1162 response with various flags. New helper functions ASN1_TIME_check()
1163 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
1164 (converts ASN1_TIME to GeneralizedTime).
1167 +) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
1168 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
1169 structure from a certificate. X509_pubkey_digest() digests the public_key
1170 contents: this is used in various key identifiers.
1173 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
1174 keyUsage if basicConstraints absent for a CA.
1177 *) Make SMIME_write_PKCS7() write mail header values with a format that
1178 is more generally accepted (no spaces before the semicolon), since
1179 some programs can't parse those values properly otherwise. Also make
1180 sure BIO's that break lines after each write do not create invalid
1184 +) Make sk_sort() tolerate a NULL argument.
1185 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
1187 +) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
1188 passed by the function are trusted implicitly. If any of them signed the
1189 response then it is assumed to be valid and is not verified.
1192 -) Make the CRL encoding routines work with empty SEQUENCE OF. The
1193 macros previously used would not encode an empty SEQUENCE OF
1194 and break the signature.
1196 [This change does not apply to 0.9.7.]
1198 *) Zero the premaster secret after deriving the master secret in
1202 +) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
1203 to data. This was previously part of the PKCS7 ASN1 code. This
1204 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
1205 [Steve Henson, reported by Kenneth R. Robinette
1206 <support@securenetterm.com>]
1208 +) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
1209 routines: without these tracing memory leaks is very painful.
1210 Fix leaks in PKCS12 and PKCS7 routines.
1213 *) Add some EVP_add_digest_alias registrations (as found in
1214 OpenSSL_add_all_digests()) to SSL_library_init()
1215 aka OpenSSL_add_ssl_algorithms(). This provides improved
1216 compatibility with peers using X.509 certificates
1217 with unconventional AlgorithmIdentifier OIDs.
1220 *) Fix for Irix with NO_ASM.
1221 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
1223 *) ./config script fixes.
1224 [Ulf Moeller, Richard Levitte]
1226 +) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
1227 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
1228 effectively meant GeneralizedTime would never be used. Now it
1229 is initialised to -1 but X509_time_adj() now has to check the value
1230 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
1231 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
1232 [Steve Henson, reported by Kenneth R. Robinette
1233 <support@securenetterm.com>]
1235 +) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
1236 result in a zero length in the ASN1_INTEGER structure which was
1237 not consistent with the structure when d2i_ASN1_INTEGER() was used
1238 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
1239 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
1240 where it did not print out a minus for negative ASN1_INTEGER.
1243 *) Fix 'openssl passwd -1'.
1246 +) Add summary printout to ocsp utility. The various functions which
1247 convert status values to strings have been renamed to:
1248 OCSP_response_status_str(), OCSP_cert_status_str() and
1249 OCSP_crl_reason_str() and are no longer static. New options
1250 to verify nonce values and to disable verification. OCSP response
1251 printout format cleaned up.
1254 +) Add additional OCSP certificate checks. These are those specified
1255 in RFC2560. This consists of two separate checks: the CA of the
1256 certificate being checked must either be the OCSP signer certificate
1257 or the issuer of the OCSP signer certificate. In the latter case the
1258 OCSP signer certificate must contain the OCSP signing extended key
1259 usage. This check is performed by attempting to match the OCSP
1260 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
1261 in the OCSP_CERTID structures of the response.
1264 +) Initial OCSP certificate verification added to OCSP_basic_verify()
1265 and related routines. This uses the standard OpenSSL certificate
1266 verify routines to perform initial checks (just CA validity) and
1267 to obtain the certificate chain. Then additional checks will be
1268 performed on the chain. Currently the root CA is checked to see
1269 if it is explicitly trusted for OCSP signing. This is used to set
1270 a root CA as a global signing root: that is any certificate that
1271 chains to that CA is an acceptable OCSP signing certificate.
1274 +) New '-extfile ...' option to 'openssl ca' for reading X.509v3
1275 extensions from a separate configuration file.
1276 As when reading extensions from the main configuration file,
1277 the '-extensions ...' option may be used for specifying the
1279 [Massimiliano Pala <madwolf@comune.modena.it>]
1281 *) Change PKCS12_key_gen_asc() so it can cope with non null
1282 terminated strings whose length is passed in the passlen
1283 parameter, for example from PEM callbacks. This was done
1284 by adding an extra length parameter to asc2uni().
1285 [Steve Henson, reported by <oddissey@samsung.co.kr>]
1287 +) New OCSP utility. Allows OCSP requests to be generated or
1288 read. The request can be sent to a responder and the output
1289 parsed, outputed or printed in text form. Not complete yet:
1290 still needs to check the OCSP response validity.
1293 +) New subcommands for 'openssl ca':
1294 'openssl ca -status <serial>' prints the status of the cert with
1295 the given serial number (according to the index file).
1296 'openssl ca -updatedb' updates the expiry status of certificates
1298 [Massimiliano Pala <madwolf@comune.modena.it>]
1300 +) New '-newreq-nodes' command option to CA.pl. This is like
1301 '-newreq', but calls 'openssl req' with the '-nodes' option
1302 so that the resulting key is not encrypted.
1303 [Damien Miller <djm@mindrot.org>]
1305 +) New configuration for the GNU Hurd.
1306 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
1308 +) Initial code to implement OCSP basic response verify. This
1309 is currently incomplete. Currently just finds the signer's
1310 certificate and verifies the signature on the response.
1313 +) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
1314 value of OPENSSLDIR. This is available via the new '-d' option
1315 to 'openssl version', and is also included in 'openssl version -a'.
1318 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
1319 call failed, free the DSA structure.
1322 +) Allowing defining memory allocation callbacks that will be given
1323 file name and line number information in additional arguments
1324 (a const char* and an int). The basic functionality remains, as
1325 well as the original possibility to just replace malloc(),
1326 realloc() and free() by functions that do not know about these
1327 additional arguments. To register and find out the current
1328 settings for extended allocation functions, the following
1329 functions are provided:
1331 CRYPTO_set_mem_ex_functions
1332 CRYPTO_set_locked_mem_ex_functions
1333 CRYPTO_get_mem_ex_functions
1334 CRYPTO_get_locked_mem_ex_functions
1336 These work the same way as CRYPTO_set_mem_functions and friends.
1337 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
1338 extended allocation function is enabled.
1339 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
1340 a conventional allocation function is enabled.
1341 [Richard Levitte, Bodo Moeller]
1343 *) Fix to uni2asc() to cope with zero length Unicode strings.
1344 These are present in some PKCS#12 files.
1347 +) Finish off removing the remaining LHASH function pointer casts.
1348 There should no longer be any prototype-casting required when using
1349 the LHASH abstraction, and any casts that remain are "bugs". See
1350 the callback types and macros at the head of lhash.h for details
1351 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
1354 +) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
1355 If /dev/[u]random devices are not available or do not return enough
1356 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
1358 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
1359 /etc/entropy will be queried once each in this sequence, quering stops
1360 when enough entropy was collected without querying more sockets.
1363 +) Change the Unix RAND_poll() variant to be able to poll several
1364 random devices, as specified by DEVRANDOM, until a sufficient amount
1365 of data has been collected. We spend at most 10 ms on each file
1366 (select timeout) and read in non-blocking mode. DEVRANDOM now
1367 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
1368 (previously it was just the string "/dev/urandom"), so on typical
1369 platforms the 10 ms delay will never occur.
1370 Also separate out the Unix variant to its own file, rand_unix.c.
1371 For VMS, there's a currently-empty rand_vms.c.
1374 +) Move OCSP client related routines to ocsp_cl.c. These
1375 provide utility functions which an application needing
1376 to issue a request to an OCSP responder and analyse the
1377 response will typically need: as opposed to those which an
1378 OCSP responder itself would need which will be added later.
1380 OCSP_request_sign() signs an OCSP request with an API similar
1381 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
1382 response. OCSP_response_get1_basic() extracts basic response
1383 from response. OCSP_resp_find_status(): finds and extracts status
1384 information from an OCSP_CERTID structure (which will be created
1385 when the request structure is built). These are built from lower
1386 level functions which work on OCSP_SINGLERESP structures but
1387 wont normally be used unless the application wishes to examine
1388 extensions in the OCSP response for example.
1390 Replace nonce routines with a pair of functions.
1391 OCSP_request_add1_nonce() adds a nonce value and optionally
1392 generates a random value. OCSP_check_nonce() checks the
1393 validity of the nonce in an OCSP response.
1396 +) Change function OCSP_request_add() to OCSP_request_add0_id().
1397 This doesn't copy the supplied OCSP_CERTID and avoids the
1398 need to free up the newly created id. Change return type
1399 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
1400 This can then be used to add extensions to the request.
1401 Deleted OCSP_request_new(), since most of its functionality
1402 is now in OCSP_REQUEST_new() (and the case insensitive name
1403 clash) apart from the ability to set the request name which
1404 will be added elsewhere.
1407 +) Update OCSP API. Remove obsolete extensions argument from
1408 various functions. Extensions are now handled using the new
1409 OCSP extension code. New simple OCSP HTTP function which
1410 can be used to send requests and parse the response.
1413 +) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
1414 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
1415 uses the special reorder version of SET OF to sort the attributes
1416 and reorder them to match the encoded order. This resolves a long
1417 standing problem: a verify on a PKCS7 structure just after signing
1418 it used to fail because the attribute order did not match the
1419 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
1420 it uses the received order. This is necessary to tolerate some broken
1421 software that does not order SET OF. This is handled by encoding
1422 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
1423 to produce the required SET OF.
1426 +) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
1427 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
1428 files to get correct declarations of the ASN.1 item variables.
1431 +) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
1432 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
1433 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
1434 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
1435 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
1436 ASN1_ITEM and no wrapper functions.
1439 +) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
1440 replace the old function pointer based I/O routines. Change most of
1441 the *_d2i_bio() and *_d2i_fp() functions to use these.
1444 +) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
1445 lines, recognice more "algorithms" that can be deselected, and make
1446 it complain about algorithm deselection that isn't recognised.
1449 +) New ASN1 functions to handle dup, sign, verify, digest, pack and
1450 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
1451 to use new functions. Add NO_ASN1_OLD which can be set to remove
1452 some old style ASN1 functions: this can be used to determine if old
1453 code will still work when these eventually go away.
1456 +) New extension functions for OCSP structures, these follow the
1457 same conventions as certificates and CRLs.
1460 +) New function X509V3_add1_i2d(). This automatically encodes and
1461 adds an extension. Its behaviour can be customised with various
1462 flags to append, replace or delete. Various wrappers added for
1463 certifcates and CRLs.
1466 +) Fix to avoid calling the underlying ASN1 print routine when
1467 an extension cannot be parsed. Correct a typo in the
1468 OCSP_SERVICELOC extension. Tidy up print OCSP format.
1471 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
1472 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
1473 when writing a 32767 byte record.
1474 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
1476 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
1477 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
1479 (RSA objects have a reference count access to which is protected
1480 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
1481 so they are meant to be shared between threads.)
1482 [Bodo Moeller, Geoff Thorpe; original patch submitted by
1483 "Reddie, Steven" <Steven.Reddie@ca.com>]
1485 +) Make mkdef.pl parse some of the ASN1 macros and add apropriate
1486 entries for variables.
1489 *) Fix a deadlock in CRYPTO_mem_leaks().
1492 +) Add functionality to apps/openssl.c for detecting locking
1493 problems: As the program is single-threaded, all we have
1494 to do is register a locking callback using an array for
1495 storing which locks are currently held by the program.
1498 +) Use a lock around the call to CRYPTO_get_ex_new_index() in
1499 SSL_get_ex_data_X509_STORE_idx(), which is used in
1500 ssl_verify_cert_chain() and thus can be called at any time
1501 during TLS/SSL handshakes so that thread-safety is essential.
1502 Unfortunately, the ex_data design is not at all suited
1503 for multi-threaded use, so it probably should be abolished.
1506 +) Added Broadcom "ubsec" ENGINE to OpenSSL.
1507 [Broadcom, tweaked and integrated by Geoff Thorpe]
1509 +) Move common extension printing code to new function
1510 X509V3_print_extensions(). Reorganise OCSP print routines and
1511 implement some needed OCSP ASN1 functions. Add OCSP extensions.
1514 +) New function X509_signature_print() to remove duplication in some
1518 +) Add a special meaning when SET OF and SEQUENCE OF flags are both
1519 set (this was treated exactly the same as SET OF previously). This
1520 is used to reorder the STACK representing the structure to match the
1521 encoding. This will be used to get round a problem where a PKCS7
1522 structure which was signed could not be verified because the STACK
1523 order did not reflect the encoded order.
1526 +) Reimplement the OCSP ASN1 module using the new code.
1529 +) Update the X509V3 code to permit the use of an ASN1_ITEM structure
1530 for its ASN1 operations. The old style function pointers still exist
1531 for now but they will eventually go away.
1534 +) Merge in replacement ASN1 code from the ASN1 branch. This almost
1535 completely replaces the old ASN1 functionality with a table driven
1536 encoder and decoder which interprets an ASN1_ITEM structure describing
1537 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
1538 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
1539 has also been converted to the new form.
1542 +) Change BN_mod_exp_recp so that negative moduli are tolerated
1543 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
1544 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
1545 for negative moduli.
1548 +) Fix BN_uadd and BN_usub: Always return non-negative results instead
1549 of not touching the result's sign bit.
1552 +) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
1556 +) Changed the LHASH code to use prototypes for callbacks, and created
1557 macros to declare and implement thin (optionally static) functions
1558 that provide type-safety and avoid function pointer casting for the
1559 type-specific callbacks.
1562 *) Use better test patterns in bntest.
1565 +) Added Kerberos Cipher Suites to be used with TLS, as written in
1567 [Veers Staats <staatsvr@asc.hpc.mil>,
1568 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
1570 *) rand_win.c fix for Borland C.
1573 *) BN_rshift bugfix for n == 0.
1576 +) Reformat the FAQ so the different questions and answers can be divided
1577 in sections depending on the subject.
1580 +) Have the zlib compression code load ZLIB.DLL dynamically under
1584 +) New function BN_mod_sqrt for computing square roots modulo a prime
1585 (using the probabilistic Tonelli-Shanks algorithm unless
1586 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
1587 be handled deterministically).
1588 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
1590 *) Add a 'bctest' script that checks for some known 'bc' bugs
1591 so that 'make test' does not abort just because 'bc' is broken.
1594 *) Store verify_result within SSL_SESSION also for client side to
1595 avoid potential security hole. (Re-used sessions on the client side
1596 always resulted in verify_result==X509_V_OK, not using the original
1597 result of the server certificate verification.)
1600 +) Make BN_mod_inverse faster by explicitly handling small quotients
1601 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
1602 512 bits], about 30% for larger ones [1024 or 2048 bits].)
1605 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
1606 SSL3_RT_APPLICATION_DATA, return 0.
1607 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
1611 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
1612 releases, have been re-implemented by renaming the previous
1613 implementations of ssl2_read and ssl3_read to ssl2_read_internal
1614 and ssl3_read_internal, respectively, and adding 'peek' parameters
1615 to them. The new ssl[23]_{read,peek} functions are calls to
1616 ssl[23]_read_internal with the 'peek' flag set appropriately.
1617 A 'peek' parameter has also been added to ssl3_read_bytes, which
1618 does the actual work for ssl3_read_internal.
1621 +) New function BN_kronecker.
1624 +) Fix BN_gcd so that it works on negative inputs; the result is
1625 positive unless both parameters are zero.
1626 Previously something reasonably close to an infinite loop was
1627 possible because numbers could be growing instead of shrinking
1628 in the implementation of Euclid's algorithm.
1631 +) Fix BN_is_word() and BN_is_one() macros to take into account the
1632 sign of the number in question.
1634 Fix BN_is_word(a,w) to work correctly for w == 0.
1636 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
1637 because its test if the absolute value of 'a' equals 'w'.
1638 Note that BN_abs_is_word does *not* handle w == 0 reliably;
1639 it exists mostly for use in the implementations of BN_is_zero(),
1640 BN_is_one(), and BN_is_word().
1643 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
1644 the method-specific "init()" handler. Also clean up ex_data after
1645 calling the method-specific "finish()" handler. Previously, this was
1646 happening the other way round.
1649 +) New function BN_swap.
1652 +) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
1653 the exponentiation functions are more likely to produce reasonable
1654 results on negative inputs.
1657 +) Change BN_mod_mul so that the result is always non-negative.
1658 Previously, it could be negative if one of the factors was negative;
1659 I don't think anyone really wanted that behaviour.
1662 +) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
1663 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
1664 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
1665 and add new functions:
1674 BN_mod_lshift1_quick
1678 These functions always generate non-negative results.
1680 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
1681 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
1683 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
1684 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
1685 be reduced modulo m.
1686 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
1688 +) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
1689 was actually never needed) and in BN_mul(). The removal in BN_mul()
1690 required a small change in bn_mul_part_recursive() and the addition
1691 of the functions bn_cmp_part_words(), bn_sub_part_words() and
1692 bn_add_part_words(), which do the same thing as bn_cmp_words(),
1693 bn_sub_words() and bn_add_words() except they take arrays with
1697 +) In 'openssl passwd', verify passwords read from the terminal
1698 unless the '-salt' option is used (which usually means that
1699 verification would just waste user's time since the resulting
1700 hash is going to be compared with some given password hash)
1701 or the new '-noverify' option is used.
1703 This is an incompatible change, but it does not affect
1704 non-interactive use of 'openssl passwd' (passwords on the command
1705 line, '-stdin' option, '-in ...' option) and thus should not
1709 +) Remove all references to RSAref, since there's no more need for it.
1712 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
1713 The previous value, 12, was not always sufficient for BN_mod_exp().
1716 +) Make DSO load along a path given through an environment variable
1717 (SHLIB_PATH) with shl_load().
1720 +) Constify the ENGINE code as a result of BIGNUM constification.
1721 Also constify the RSA code and most things related to it. In a
1722 few places, most notable in the depth of the ASN.1 code, ugly
1723 casts back to non-const were required (to be solved at a later
1727 +) Make it so the openssl application has all engines loaded by default.
1730 +) Constify the BIGNUM routines a little more.
1733 *) Make sure that shared libraries get the internal name engine with
1734 the full version number and not just 0. This should mark the
1735 shared libraries as not backward compatible. Of course, this should
1736 be changed again when we can guarantee backward binary compatibility.
1739 +) Add the following functions:
1741 ENGINE_load_cswift()
1743 ENGINE_load_atalla()
1745 ENGINE_load_builtin_engines()
1747 That way, an application can itself choose if external engines that
1748 are built-in in OpenSSL shall ever be used or not. The benefit is
1749 that applications won't have to be linked with libdl or other dso
1750 libraries unless it's really needed.
1752 Changed 'openssl engine' to load all engines on demand.
1753 Changed the engine header files to avoid the duplication of some
1754 declarations (they differed!).
1757 +) 'openssl engine' can now list capabilities.
1760 +) Better error reporting in 'openssl engine'.
1763 +) Never call load_dh_param(NULL) in s_server.
1766 +) Add engine application. It can currently list engines by name and
1767 identity, and test if they are actually available.
1770 +) Add support for shared libraries under Irix.
1771 [Albert Chin-A-Young <china@thewrittenword.com>]
1773 +) Improve RPM specification file by forcing symbolic linking and making
1774 sure the installed documentation is also owned by root.root.
1775 [Damien Miller <djm@mindrot.org>]
1777 +) Add configuration option to build on Linux on both big-endian and
1779 [Ralf Baechle <ralf@uni-koblenz.de>]
1781 +) Give the OpenSSL applications more possibilities to make use of
1782 keys (public as well as private) handled by engines.
1785 +) Add OCSP code that comes from CertCo.
1788 +) Add VMS support for the Rijndael code.
1791 +) Added untested support for Nuron crypto accelerator.
1794 +) Add support for external cryptographic devices. This code was
1795 previously distributed separately as the "engine" branch.
1796 [Geoff Thorpe, Richard Levitte]
1798 +) Rework the filename-translation in the DSO code. It is now possible to
1799 have far greater control over how a "name" is turned into a filename
1800 depending on the operating environment and any oddities about the
1801 different shared library filenames on each system.
1804 +) Support threads on FreeBSD-elf in Configure.
1807 +) Add the possibility to create shared libraries on HP-UX
1810 +) Fix for SHA1 assembly problem with MASM: it produces
1811 warnings about corrupt line number information when assembling
1812 with debugging information. This is caused by the overlapping
1814 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
1817 NCONF_get_number() has no error checking at all. As a replacement,
1818 NCONF_get_number_e() is defined (_e for "error checking") and is
1819 promoted strongly. The old NCONF_get_number is kept around for
1820 binary backward compatibility.
1821 Make it possible for methods to load from something other than a BIO,
1822 by providing a function pointer that is given a name instead of a BIO.
1823 For example, this could be used to load configuration data from an
1827 *) Fix typo in get_cert_by_subject() in by_dir.c
1828 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
1830 *) Rework the system to generate shared libraries:
1832 - Make note of the expected extension for the shared libraries and
1833 if there is a need for symbolic links from for example libcrypto.so.0
1834 to libcrypto.so.0.9.7. There is extended info in Configure for
1837 - Make as few rebuilds of the shared libraries as possible.
1839 - Still avoid linking the OpenSSL programs with the shared libraries.
1841 - When installing, install the shared libraries separately from the
1845 +) Fix for non blocking accept BIOs. Added new I/O special reason
1846 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
1847 with non blocking I/O was not possible because no retry code was
1848 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
1852 +) Added the beginnings of Rijndael support.
1855 +) Fix for bug in DirectoryString mask setting. Add support for
1856 X509_NAME_print_ex() in 'req' and X509_print_ex() function
1857 to allow certificate printing to more controllable, additional
1858 'certopt' option to 'x509' to allow new printing options to be
1862 +) Clean old EAY MD5 hack from e_os.h.
1865 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
1867 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
1868 and not in SSL_clear because the latter is also used by the
1869 accept/connect functions; previously, the settings made by
1870 SSL_set_read_ahead would be lost during the handshake.
1871 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
1873 *) Correct util/mkdef.pl to be selective about disabled algorithms.
1874 Previously, it would create entries for disableed algorithms no
1878 *) Added several new manual pages for SSL_* function.
1881 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
1883 *) In ssl23_get_client_hello, generate an error message when faced
1884 with an initial SSL 3.0/TLS record that is too small to contain the
1885 first two bytes of the ClientHello message, i.e. client_version.
1886 (Note that this is a pathologic case that probably has never happened
1887 in real life.) The previous approach was to use the version number
1888 from the record header as a substitute; but our protocol choice
1889 should not depend on that one because it is not authenticated
1890 by the Finished messages.
1893 *) More robust randomness gathering functions for Windows.
1894 [Jeffrey Altman <jaltman@columbia.edu>]
1896 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
1897 not set then we don't setup the error code for issuer check errors
1898 to avoid possibly overwriting other errors which the callback does
1899 handle. If an application does set the flag then we assume it knows
1900 what it is doing and can handle the new informational codes
1904 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
1905 a general "ANY" type, as such it should be able to decode anything
1906 including tagged types. However it didn't check the class so it would
1907 wrongly interpret tagged types in the same way as their universal
1908 counterpart and unknown types were just rejected. Changed so that the
1909 tagged and unknown types are handled in the same way as a SEQUENCE:
1910 that is the encoding is stored intact. There is also a new type
1911 "V_ASN1_OTHER" which is used when the class is not universal, in this
1912 case we have no idea what the actual type is so we just lump them all
1916 *) On VMS, stdout may very well lead to a file that is written to
1917 in a record-oriented fashion. That means that every write() will
1918 write a separate record, which will be read separately by the
1919 programs trying to read from it. This can be very confusing.
1921 The solution is to put a BIO filter in the way that will buffer
1922 text until a linefeed is reached, and then write everything a
1923 line at a time, so every record written will be an actual line,
1924 not chunks of lines and not (usually doesn't happen, but I've
1925 seen it once) several lines in one record. BIO_f_linebuffer() is
1928 Currently, it's a VMS-only method, because that's where it has
1929 been tested well enough.
1932 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
1933 it can return incorrect results.
1934 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
1935 but it was in 0.9.6-beta[12].)
1938 *) Disable the check for content being present when verifying detached
1939 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
1940 include zero length content when signing messages.
1943 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
1944 BIO_ctrl (for BIO pairs).
1947 *) Add DSO method for VMS.
1950 *) Bug fix: Montgomery multiplication could produce results with the
1954 *) Add RPM specification openssl.spec and modify it to build three
1955 packages. The default package contains applications, application
1956 documentation and run-time libraries. The devel package contains
1957 include files, static libraries and function documentation. The
1958 doc package contains the contents of the doc directory. The original
1959 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
1962 *) Add a large number of documentation files for many SSL routines.
1963 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
1965 *) Add a configuration entry for Sony News 4.
1966 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
1968 *) Don't set the two most significant bits to one when generating a
1969 random number < q in the DSA library.
1972 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
1973 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
1974 the underlying transport is blocking) if a handshake took place.
1975 (The default behaviour is needed by applications such as s_client
1976 and s_server that use select() to determine when to use SSL_read;
1977 but for applications that know in advance when to expect data, it
1978 just makes things more complicated.)
1981 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
1985 *) Add a few more EBCDIC conditionals that make `req' and `x509'
1986 work better on such systems.
1987 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
1989 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
1990 Update PKCS12_parse() so it copies the friendlyName and the
1991 keyid to the certificates aux info.
1994 *) Fix bug in PKCS7_verify() which caused an infinite loop
1995 if there was more than one signature.
1996 [Sven Uszpelkat <su@celocom.de>]
1998 *) Major change in util/mkdef.pl to include extra information
1999 about each symbol, as well as presentig variables as well
2000 as functions. This change means that there's n more need
2001 to rebuild the .num files when some algorithms are excluded.
2004 *) Allow the verify time to be set by an application,
2005 rather than always using the current time.
2008 *) Phase 2 verify code reorganisation. The certificate
2009 verify code now looks up an issuer certificate by a
2010 number of criteria: subject name, authority key id
2011 and key usage. It also verifies self signed certificates
2012 by the same criteria. The main comparison function is
2013 X509_check_issued() which performs these checks.
2015 Lot of changes were necessary in order to support this
2016 without completely rewriting the lookup code.
2018 Authority and subject key identifier are now cached.
2020 The LHASH 'certs' is X509_STORE has now been replaced
2021 by a STACK_OF(X509_OBJECT). This is mainly because an
2022 LHASH can't store or retrieve multiple objects with
2023 the same hash value.
2025 As a result various functions (which were all internal
2026 use only) have changed to handle the new X509_STORE
2027 structure. This will break anything that messed round
2028 with X509_STORE internally.
2030 The functions X509_STORE_add_cert() now checks for an
2031 exact match, rather than just subject name.
2033 The X509_STORE API doesn't directly support the retrieval
2034 of multiple certificates matching a given criteria, however
2035 this can be worked round by performing a lookup first
2036 (which will fill the cache with candidate certificates)
2037 and then examining the cache for matches. This is probably
2038 the best we can do without throwing out X509_LOOKUP
2039 entirely (maybe later...).
2041 The X509_VERIFY_CTX structure has been enhanced considerably.
2043 All certificate lookup operations now go via a get_issuer()
2044 callback. Although this currently uses an X509_STORE it
2045 can be replaced by custom lookups. This is a simple way
2046 to bypass the X509_STORE hackery necessary to make this
2047 work and makes it possible to use more efficient techniques
2048 in future. A very simple version which uses a simple
2049 STACK for its trusted certificate store is also provided
2050 using X509_STORE_CTX_trusted_stack().
2052 The verify_cb() and verify() callbacks now have equivalents
2053 in the X509_STORE_CTX structure.
2055 X509_STORE_CTX also has a 'flags' field which can be used
2056 to customise the verify behaviour.
2059 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
2060 excludes S/MIME capabilities.
2063 *) When a certificate request is read in keep a copy of the
2064 original encoding of the signed data and use it when outputing
2065 again. Signatures then use the original encoding rather than
2066 a decoded, encoded version which may cause problems if the
2067 request is improperly encoded.
2070 *) For consistency with other BIO_puts implementations, call
2071 buffer_write(b, ...) directly in buffer_puts instead of calling
2074 In BIO_puts, increment b->num_write as in BIO_write.
2075 [Peter.Sylvester@EdelWeb.fr]
2077 *) Fix BN_mul_word for the case where the word is 0. (We have to use
2078 BN_zero, we may not return a BIGNUM with an array consisting of
2082 *) Avoid calling abort() from within the library when problems are
2083 detected, except if preprocessor symbols have been defined
2084 (such as REF_CHECK, BN_DEBUG etc.).
2087 *) New openssl application 'rsautl'. This utility can be
2088 used for low level RSA operations. DER public key
2089 BIO/fp routines also added.
2092 *) New Configure entry and patches for compiling on QNX 4.
2093 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
2095 *) A demo state-machine implementation was sponsored by
2096 Nuron (http://www.nuron.com/) and is now available in
2097 demos/state_machine.
2100 *) New options added to the 'dgst' utility for signature
2101 generation and verification.
2104 *) Unrecognized PKCS#7 content types are now handled via a
2105 catch all ASN1_TYPE structure. This allows unsupported
2106 types to be stored as a "blob" and an application can
2107 encode and decode it manually.
2110 *) Fix various signed/unsigned issues to make a_strex.c
2112 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
2114 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
2115 length if passed a buffer. ASN1_INTEGER_to_BN failed
2116 if passed a NULL BN and its argument was negative.
2117 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
2119 *) Modification to PKCS#7 encoding routines to output definite
2120 length encoding. Since currently the whole structures are in
2121 memory there's not real point in using indefinite length
2122 constructed encoding. However if OpenSSL is compiled with
2123 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
2126 *) Added BIO_vprintf() and BIO_vsnprintf().
2129 *) Added more prefixes to parse for in the the strings written
2130 through a logging bio, to cover all the levels that are available
2131 through syslog. The prefixes are now:
2133 PANIC, EMERG, EMR => LOG_EMERG
2134 ALERT, ALR => LOG_ALERT
2135 CRIT, CRI => LOG_CRIT
2136 ERROR, ERR => LOG_ERR
2137 WARNING, WARN, WAR => LOG_WARNING
2138 NOTICE, NOTE, NOT => LOG_NOTICE
2139 INFO, INF => LOG_INFO
2140 DEBUG, DBG => LOG_DEBUG
2142 and as before, if none of those prefixes are present at the
2143 beginning of the string, LOG_ERR is chosen.
2145 On Win32, the LOG_* levels are mapped according to this:
2147 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
2148 LOG_WARNING => EVENTLOG_WARNING_TYPE
2149 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
2153 *) Made it possible to reconfigure with just the configuration
2154 argument "reconf" or "reconfigure". The command line arguments
2155 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
2156 and are retrieved from there when reconfiguring.
2160 [Assar Westerlund <assar@sics.se>, Richard Levitte]
2162 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
2165 *) The obj_dat.pl script was messing up the sorting of object
2166 names. The reason was that it compared the quoted version
2167 of strings as a result "OCSP" > "OCSP Signing" because
2168 " > SPACE. Changed script to store unquoted versions of
2169 names and add quotes on output. It was also omitting some
2170 names from the lookup table if they were given a default
2171 value (that is if SN is missing it is given the same
2172 value as LN and vice versa), these are now added on the
2173 grounds that if an object has a name we should be able to
2174 look it up. Finally added warning output when duplicate
2175 short or long names are found.
2178 *) Changes needed for Tandem NSK.
2179 [Scott Uroff <scott@xypro.com>]
2181 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
2182 RSA_padding_check_SSLv23(), special padding was never detected
2183 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
2184 version rollback attacks was not effective.
2186 In s23_clnt.c, don't use special rollback-attack detection padding
2187 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
2188 client; similarly, in s23_srvr.c, don't do the rollback check if
2189 SSL 2.0 is the only protocol enabled in the server.
2192 *) Make it possible to get hexdumps of unprintable data with 'openssl
2193 asn1parse'. By implication, the functions ASN1_parse_dump() and
2194 BIO_dump_indent() are added.
2197 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
2198 these print out strings and name structures based on various
2199 flags including RFC2253 support and proper handling of
2200 multibyte characters. Added options to the 'x509' utility
2201 to allow the various flags to be set.
2204 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
2205 Also change the functions X509_cmp_current_time() and
2206 X509_gmtime_adj() work with an ASN1_TIME structure,
2207 this will enable certificates using GeneralizedTime in validity
2208 dates to be checked.
2211 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
2212 negative public key encodings) on by default,
2213 NO_NEG_PUBKEY_BUG can be set to disable it.
2216 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
2217 content octets. An i2c_ASN1_OBJECT is unnecessary because
2218 the encoding can be trivially obtained from the structure.
2221 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
2222 not read locks (CRYPTO_r_[un]lock).
2225 *) A first attempt at creating official support for shared
2226 libraries through configuration. I've kept it so the
2227 default is static libraries only, and the OpenSSL programs
2228 are always statically linked for now, but there are
2229 preparations for dynamic linking in place.
2230 This has been tested on Linux and Tru64.
2233 *) Randomness polling function for Win9x, as described in:
2234 Peter Gutmann, Software Generation of Practically Strong
2238 *) Fix so PRNG is seeded in req if using an already existing
2242 *) New options to smime application. -inform and -outform
2243 allow alternative formats for the S/MIME message including
2244 PEM and DER. The -content option allows the content to be
2245 specified separately. This should allow things like Netscape
2246 form signing output easier to verify.
2249 *) Fix the ASN1 encoding of tags using the 'long form'.
2252 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
2253 STRING types. These convert content octets to and from the
2254 underlying type. The actual tag and length octets are
2255 already assumed to have been read in and checked. These
2256 are needed because all other string types have virtually
2257 identical handling apart from the tag. By having versions
2258 of the ASN1 functions that just operate on content octets
2259 IMPLICIT tagging can be handled properly. It also allows
2260 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
2261 and ASN1_INTEGER are identical apart from the tag.
2264 *) Change the handling of OID objects as follows:
2266 - New object identifiers are inserted in objects.txt, following
2267 the syntax given in objects.README.
2268 - objects.pl is used to process obj_mac.num and create a new
2270 - obj_dat.pl is used to create a new obj_dat.h, using the data in
2273 This is currently kind of a hack, and the perl code in objects.pl
2274 isn't very elegant, but it works as I intended. The simplest way
2275 to check that it worked correctly is to look in obj_dat.h and
2276 check the array nid_objs and make sure the objects haven't moved
2277 around (this is important!). Additions are OK, as well as
2278 consistent name changes.
2281 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
2284 *) Addition of the command line parameter '-rand file' to 'openssl req'.
2285 The given file adds to whatever has already been seeded into the
2286 random pool through the RANDFILE configuration file option or
2287 environment variable, or the default random state file.
2290 *) mkstack.pl now sorts each macro group into lexical order.
2291 Previously the output order depended on the order the files
2292 appeared in the directory, resulting in needless rewriting
2296 *) Patches to make OpenSSL compile under Win32 again. Mostly
2297 work arounds for the VC++ problem that it treats func() as
2298 func(void). Also stripped out the parts of mkdef.pl that
2299 added extra typesafe functions: these no longer exist.
2302 *) Reorganisation of the stack code. The macros are now all
2303 collected in safestack.h . Each macro is defined in terms of
2304 a "stack macro" of the form SKM_<name>(type, a, b). The
2305 DEBUG_SAFESTACK is now handled in terms of function casts,
2306 this has the advantage of retaining type safety without the
2307 use of additional functions. If DEBUG_SAFESTACK is not defined
2308 then the non typesafe macros are used instead. Also modified the
2309 mkstack.pl script to handle the new form. Needs testing to see
2310 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
2311 the default if no major problems. Similar behaviour for ASN1_SET_OF
2312 and PKCS12_STACK_OF.
2315 *) When some versions of IIS use the 'NET' form of private key the
2316 key derivation algorithm is different. Normally MD5(password) is
2317 used as a 128 bit RC4 key. In the modified case
2318 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
2319 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
2320 as the old Netscape_RSA functions except they have an additional
2321 'sgckey' parameter which uses the modified algorithm. Also added
2322 an -sgckey command line option to the rsa utility. Thanks to
2323 Adrian Peck <bertie@ncipher.com> for posting details of the modified
2324 algorithm to openssl-dev.
2327 *) The evp_local.h macros were using 'c.##kname' which resulted in
2328 invalid expansion on some systems (SCO 5.0.5 for example).
2329 Corrected to 'c.kname'.
2330 [Phillip Porch <root@theporch.com>]
2332 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
2333 a STACK of email addresses from a certificate or request, these look
2334 in the subject name and the subject alternative name extensions and
2335 omit any duplicate addresses.
2338 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
2339 This makes DSA verification about 2 % faster.
2342 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
2343 (meaning that now 2^5 values will be precomputed, which is only 4 KB
2344 plus overhead for 1024 bit moduli).
2345 This makes exponentiations about 0.5 % faster for 1024 bit
2346 exponents (as measured by "openssl speed rsa2048").
2349 *) Rename memory handling macros to avoid conflicts with other
2351 Malloc => OPENSSL_malloc
2352 Malloc_locked => OPENSSL_malloc_locked
2353 Realloc => OPENSSL_realloc
2354 Free => OPENSSL_free
2357 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
2358 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
2361 *) CygWin32 support.
2362 [John Jarvie <jjarvie@newsguy.com>]
2364 *) The type-safe stack code has been rejigged. It is now only compiled
2365 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
2366 by default all type-specific stack functions are "#define"d back to
2367 standard stack functions. This results in more streamlined output
2368 but retains the type-safety checking possibilities of the original
2372 *) The STACK code has been cleaned up, and certain type declarations
2373 that didn't make a lot of sense have been brought in line. This has
2374 also involved a cleanup of sorts in safestack.h to more correctly
2375 map type-safe stack functions onto their plain stack counterparts.
2376 This work has also resulted in a variety of "const"ifications of
2377 lots of the code, especially "_cmp" operations which should normally
2378 be prototyped with "const" parameters anyway.
2381 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
2382 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
2383 (The PRNG state consists of two parts, the large pool 'state' and 'md',
2384 where all of 'md' is used each time the PRNG is used, but 'state'
2385 is used only indexed by a cyclic counter. As entropy may not be
2386 well distributed from the beginning, 'md' is important as a
2387 chaining variable. However, the output function chains only half
2388 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
2389 all of 'md', and seeding with STATE_SIZE dummy bytes will result
2390 in all of 'state' being rewritten, with the new values depending
2391 on virtually all of 'md'. This overcomes the 80 bit limitation.)
2394 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
2395 the handshake is continued after ssl_verify_cert_chain();
2396 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
2397 can lead to 'unexplainable' connection aborts later.
2398 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
2400 *) Major EVP API cipher revision.
2401 Add hooks for extra EVP features. This allows various cipher
2402 parameters to be set in the EVP interface. Support added for variable
2403 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
2404 setting of RC2 and RC5 parameters.
2406 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
2409 Remove lots of duplicated code from the EVP library. For example *every*
2410 cipher init() function handles the 'iv' in the same way according to the
2411 cipher mode. They also all do nothing if the 'key' parameter is NULL and
2412 for CFB and OFB modes they zero ctx->num.
2414 New functionality allows removal of S/MIME code RC2 hack.
2416 Most of the routines have the same form and so can be declared in terms
2419 By shifting this to the top level EVP_CipherInit() it can be removed from
2420 all individual ciphers. If the cipher wants to handle IVs or keys
2421 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
2424 Change lots of functions like EVP_EncryptUpdate() to now return a
2425 value: although software versions of the algorithms cannot fail
2426 any installed hardware versions can.
2429 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
2430 this option is set, tolerate broken clients that send the negotiated
2431 protocol version number instead of the requested protocol version
2435 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
2436 i.e. non-zero for export ciphersuites, zero otherwise.
2437 Previous versions had this flag inverted, inconsistent with
2438 rsa_tmp_cb (..._TMP_RSA_CB).
2439 [Bodo Moeller; problem reported by Amit Chopra]
2441 *) Add missing DSA library text string. Work around for some IIS
2442 key files with invalid SEQUENCE encoding.
2445 *) Add a document (doc/standards.txt) that list all kinds of standards
2446 and so on that are implemented in OpenSSL.
2449 *) Enhance c_rehash script. Old version would mishandle certificates
2450 with the same subject name hash and wouldn't handle CRLs at all.
2451 Added -fingerprint option to crl utility, to support new c_rehash
2455 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
2458 *) Fix for SSL server purpose checking. Server checking was
2459 rejecting certificates which had extended key usage present
2460 but no ssl client purpose.
2461 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
2463 *) Make PKCS#12 code work with no password. The PKCS#12 spec
2464 is a little unclear about how a blank password is handled.
2465 Since the password in encoded as a BMPString with terminating
2466 double NULL a zero length password would end up as just the
2467 double NULL. However no password at all is different and is
2468 handled differently in the PKCS#12 key generation code. NS
2469 treats a blank password as zero length. MSIE treats it as no
2470 password on export: but it will try both on import. We now do
2471 the same: PKCS12_parse() tries zero length and no password if
2472 the password is set to "" or NULL (NULL is now a valid password:
2473 it wasn't before) as does the pkcs12 application.
2476 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
2477 perror when PEM_read_bio_X509_REQ fails, the error message must
2478 be obtained from the error queue.
2481 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
2482 it in ERR_remove_state if appropriate, and change ERR_get_state
2483 accordingly to avoid race conditions (this is necessary because
2484 thread_hash is no longer constant once set).
2487 *) Bugfix for linux-elf makefile.one.
2490 *) RSA_get_default_method() will now cause a default
2491 RSA_METHOD to be chosen if one doesn't exist already.
2492 Previously this was only set during a call to RSA_new()
2493 or RSA_new_method(NULL) meaning it was possible for
2494 RSA_get_default_method() to return NULL.
2497 *) Added native name translation to the existing DSO code
2498 that will convert (if the flag to do so is set) filenames
2499 that are sufficiently small and have no path information
2500 into a canonical native form. Eg. "blah" converted to
2501 "libblah.so" or "blah.dll" etc.
2504 *) New function ERR_error_string_n(e, buf, len) which is like
2505 ERR_error_string(e, buf), but writes at most 'len' bytes
2506 including the 0 terminator. For ERR_error_string_n, 'buf'
2508 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
2510 *) CONF library reworked to become more general. A new CONF
2511 configuration file reader "class" is implemented as well as a
2512 new functions (NCONF_*, for "New CONF") to handle it. The now
2513 old CONF_* functions are still there, but are reimplemented to
2514 work in terms of the new functions. Also, a set of functions
2515 to handle the internal storage of the configuration data is
2516 provided to make it easier to write new configuration file
2517 reader "classes" (I can definitely see something reading a
2518 configuration file in XML format, for example), called _CONF_*,
2519 or "the configuration storage API"...
2521 The new configuration file reading functions are:
2523 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
2524 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
2526 NCONF_default, NCONF_WIN32
2528 NCONF_dump_fp, NCONF_dump_bio
2530 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
2531 NCONF_new creates a new CONF object. This works in the same way
2532 as other interfaces in OpenSSL, like the BIO interface.
2533 NCONF_dump_* dump the internal storage of the configuration file,
2534 which is useful for debugging. All other functions take the same
2535 arguments as the old CONF_* functions wth the exception of the
2536 first that must be a `CONF *' instead of a `LHASH *'.
2538 To make it easer to use the new classes with the old CONF_* functions,
2539 the function CONF_set_default_method is provided.
2542 *) Add '-tls1' option to 'openssl ciphers', which was already
2543 mentioned in the documentation but had not been implemented.
2544 (This option is not yet really useful because even the additional
2545 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
2548 *) Initial DSO code added into libcrypto for letting OpenSSL (and
2549 OpenSSL-based applications) load shared libraries and bind to
2550 them in a portable way.
2551 [Geoff Thorpe, with contributions from Richard Levitte]
2553 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
2555 *) Make sure _lrotl and _lrotr are only used with MSVC.
2557 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
2558 (the default implementation of RAND_status).
2560 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
2561 to '-clrext' (= clear extensions), as intended and documented.
2562 [Bodo Moeller; inconsistency pointed out by Michael Attili
2565 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
2566 was larger than the MD block size.
2567 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
2569 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
2570 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
2571 using the passed key: if the passed key was a private key the result
2572 of X509_print(), for example, would be to print out all the private key
2576 *) des_quad_cksum() byte order bug fix.
2577 [Ulf Möller, using the problem description in krb4-0.9.7, where
2578 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
2580 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
2582 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
2584 *) For easily testing in shell scripts whether some command
2585 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
2586 returns with exit code 0 iff no command of the given name is available.
2587 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
2588 the output goes to stdout and nothing is printed to stderr.
2589 Additional arguments are always ignored.
2591 Since for each cipher there is a command of the same name,
2592 the 'no-cipher' compilation switches can be tested this way.
2594 ('openssl no-XXX' is not able to detect pseudo-commands such
2595 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
2598 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
2601 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
2602 is set; it will be thrown away anyway because each handshake creates
2604 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
2605 to parameters -- in previous versions (since OpenSSL 0.9.3) the
2606 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
2607 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
2610 *) New s_client option -ign_eof: EOF at stdin is ignored, and
2611 'Q' and 'R' lose their special meanings (quit/renegotiate).
2612 This is part of what -quiet does; unlike -quiet, -ign_eof
2613 does not suppress any output.
2616 *) Add compatibility options to the purpose and trust code. The
2617 purpose X509_PURPOSE_ANY is "any purpose" which automatically
2618 accepts a certificate or CA, this was the previous behaviour,
2619 with all the associated security issues.
2621 X509_TRUST_COMPAT is the old trust behaviour: only and
2622 automatically trust self signed roots in certificate store. A
2623 new trust setting X509_TRUST_DEFAULT is used to specify that
2624 a purpose has no associated trust setting and it should instead
2625 use the value in the default purpose.
2628 *) Fix the PKCS#8 DSA private key code so it decodes keys again
2629 and fix a memory leak.
2632 *) In util/mkerr.pl (which implements 'make errors'), preserve
2633 reason strings from the previous version of the .c file, as
2634 the default to have only downcase letters (and digits) in
2635 automatically generated reasons codes is not always appropriate.
2638 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
2639 using strerror. Previously, ERR_reason_error_string() returned
2640 library names as reason strings for SYSerr; but SYSerr is a special
2641 case where small numbers are errno values, not library numbers.
2644 *) Add '-dsaparam' option to 'openssl dhparam' application. This
2645 converts DSA parameters into DH parameters. (When creating parameters,
2646 DSA_generate_parameters is used.)
2649 *) Include 'length' (recommended exponent length) in C code generated
2650 by 'openssl dhparam -C'.
2653 *) The second argument to set_label in perlasm was already being used
2654 so couldn't be used as a "file scope" flag. Moved to third argument
2658 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
2659 instead of RAND_bytes for encryption IVs and salts.
2662 *) Include RAND_status() into RAND_METHOD instead of implementing
2663 it only for md_rand.c Otherwise replacing the PRNG by calling
2664 RAND_set_rand_method would be impossible.
2667 *) Don't let DSA_generate_key() enter an infinite loop if the random
2668 number generation fails.
2671 *) New 'rand' application for creating pseudo-random output.
2674 *) Added configuration support for Linux/IA64
2675 [Rolf Haberrecker <rolf@suse.de>]
2677 *) Assembler module support for Mingw32.
2680 *) Shared library support for HPUX (in shlib/).
2681 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
2683 *) Shared library support for Solaris gcc.
2684 [Lutz Behnke <behnke@trustcenter.de>]
2686 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
2688 *) PKCS7_encrypt() was adding text MIME headers twice because they
2689 were added manually and by SMIME_crlf_copy().
2692 *) In bntest.c don't call BN_rand with zero bits argument.
2693 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
2695 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
2696 case was implemented. This caused BN_div_recp() to fail occasionally.
2699 *) Add an optional second argument to the set_label() in the perl
2700 assembly language builder. If this argument exists and is set
2701 to 1 it signals that the assembler should use a symbol whose
2702 scope is the entire file, not just the current function. This
2703 is needed with MASM which uses the format label:: for this scope.
2704 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
2706 *) Change the ASN1 types so they are typedefs by default. Before
2707 almost all types were #define'd to ASN1_STRING which was causing
2708 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
2712 *) Change names of new functions to the new get1/get0 naming
2713 convention: After 'get1', the caller owns a reference count
2714 and has to call ..._free; 'get0' returns a pointer to some
2715 data structure without incrementing reference counters.
2716 (Some of the existing 'get' functions increment a reference
2717 counter, some don't.)
2718 Similarly, 'set1' and 'add1' functions increase reference
2719 counters or duplicate objects.
2722 *) Allow for the possibility of temp RSA key generation failure:
2723 the code used to assume it always worked and crashed on failure.
2726 *) Fix potential buffer overrun problem in BIO_printf().
2727 [Ulf Möller, using public domain code by Patrick Powell; problem
2728 pointed out by David Sacerdote <das33@cornell.edu>]
2730 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
2731 RAND_egd() and RAND_status(). In the command line application,
2732 the EGD socket can be specified like a seed file using RANDFILE
2736 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
2737 Some CAs (e.g. Verisign) distribute certificates in this form.
2740 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
2741 list to exclude them. This means that no special compilation option
2742 is needed to use anonymous DH: it just needs to be included in the
2746 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
2747 EVP_MD_type. The old functionality is available in a new macro called
2748 EVP_MD_md(). Change code that uses it and update docs.
2751 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
2752 where the 'void *' argument is replaced by a function pointer argument.
2753 Previously 'void *' was abused to point to functions, which works on
2754 many platforms, but is not correct. As these functions are usually
2755 called by macros defined in OpenSSL header files, most source code
2756 should work without changes.
2759 *) <openssl/opensslconf.h> (which is created by Configure) now contains
2760 sections with information on -D... compiler switches used for
2761 compiling the library so that applications can see them. To enable
2762 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
2763 must be defined. E.g.,
2764 #define OPENSSL_ALGORITHM_DEFINES
2765 #include <openssl/opensslconf.h>
2766 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
2767 [Richard Levitte, Ulf and Bodo Möller]
2769 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
2773 *) Change the 'other' type in certificate aux info to a STACK_OF
2774 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
2775 the required ASN1 format: arbitrary types determined by an OID.
2778 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
2779 argument to 'req'. This is not because the function is newer or
2780 better than others it just uses the work 'NEW' in the certificate
2781 request header lines. Some software needs this.
2784 *) Reorganise password command line arguments: now passwords can be
2785 obtained from various sources. Delete the PEM_cb function and make
2786 it the default behaviour: i.e. if the callback is NULL and the
2787 usrdata argument is not NULL interpret it as a null terminated pass
2788 phrase. If usrdata and the callback are NULL then the pass phrase
2789 is prompted for as usual.
2792 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
2793 the support is automatically enabled. The resulting binaries will
2794 autodetect the card and use it if present.
2795 [Ben Laurie and Compaq Inc.]
2797 *) Work around for Netscape hang bug. This sends certificate request
2798 and server done in one record. Since this is perfectly legal in the
2799 SSL/TLS protocol it isn't a "bug" option and is on by default. See
2800 the bugs/SSLv3 entry for more info.
2803 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
2806 *) Add -rand argument to smime and pkcs12 applications and read/write
2810 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
2813 *) Add command line password options to the remaining applications.
2816 *) Bug fix for BN_div_recp() for numerators with an even number of
2820 *) More tests in bntest.c, and changed test_bn output.
2823 *) ./config recognizes MacOS X now.
2826 *) Bug fix for BN_div() when the first words of num and divsor are
2827 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
2830 *) Add support for various broken PKCS#8 formats, and command line
2831 options to produce them.
2834 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
2835 get temporary BIGNUMs from a BN_CTX.
2838 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
2842 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
2843 include a #define from the old name to the new. The original intent
2844 was that statically linked binaries could for example just call
2845 SSLeay_add_all_ciphers() to just add ciphers to the table and not
2846 link with digests. This never worked becayse SSLeay_add_all_digests()
2847 and SSLeay_add_all_ciphers() were in the same source file so calling
2848 one would link with the other. They are now in separate source files.
2851 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
2854 *) Use a less unusual form of the Miller-Rabin primality test (it used
2855 a binary algorithm for exponentiation integrated into the Miller-Rabin
2856 loop, our standard modexp algorithms are faster).
2859 *) Support for the EBCDIC character set completed.
2860 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
2862 *) Source code cleanups: use const where appropriate, eliminate casts,
2863 use void * instead of char * in lhash.
2866 *) Bugfix: ssl3_send_server_key_exchange was not restartable
2867 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
2868 this the server could overwrite ephemeral keys that the client
2872 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
2873 using 50 iterations of the Rabin-Miller test.
2875 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
2876 iterations of the Rabin-Miller test as required by the appendix
2877 to FIPS PUB 186[-1]) instead of DSA_is_prime.
2878 As BN_is_prime_fasttest includes trial division, DSA parameter
2879 generation becomes much faster.
2881 This implies a change for the callback functions in DSA_is_prime
2882 and DSA_generate_parameters: The callback function is called once
2883 for each positive witness in the Rabin-Miller test, not just
2884 occasionally in the inner loop; and the parameters to the
2885 callback function now provide an iteration count for the outer
2886 loop rather than for the current invocation of the inner loop.
2887 DSA_generate_parameters additionally can call the callback
2888 function with an 'iteration count' of -1, meaning that a
2889 candidate has passed the trial division test (when q is generated
2890 from an application-provided seed, trial division is skipped).
2893 *) New function BN_is_prime_fasttest that optionally does trial
2894 division before starting the Rabin-Miller test and has
2895 an additional BN_CTX * argument (whereas BN_is_prime always
2896 has to allocate at least one BN_CTX).
2897 'callback(1, -1, cb_arg)' is called when a number has passed the
2898 trial division stage.
2901 *) Fix for bug in CRL encoding. The validity dates weren't being handled
2905 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
2908 *) New function BN_pseudo_rand().
2911 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
2912 bignum version of BN_from_montgomery() with the working code from
2913 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
2917 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
2918 made it impossible to use the same SSL_SESSION data structure in
2919 SSL2 clients in multiple threads.
2922 *) The return value of RAND_load_file() no longer counts bytes obtained
2923 by stat(). RAND_load_file(..., -1) is new and uses the complete file
2924 to seed the PRNG (previously an explicit byte count was required).
2925 [Ulf Möller, Bodo Möller]
2927 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
2928 used (char *) instead of (void *) and had casts all over the place.
2931 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
2934 *) Retain source code compatibility for BN_prime_checks macro:
2935 BN_is_prime(..., BN_prime_checks, ...) now uses
2936 BN_prime_checks_for_size to determine the appropriate number of
2937 Rabin-Miller iterations.
2940 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
2941 DH_CHECK_P_NOT_SAFE_PRIME.
2942 (Check if this is true? OpenPGP calls them "strong".)
2945 *) Merge the functionality of "dh" and "gendh" programs into a new program
2946 "dhparam". The old programs are retained for now but will handle DH keys
2947 (instead of parameters) in future.
2950 *) Make the ciphers, s_server and s_client programs check the return values
2951 when a new cipher list is set.
2954 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
2955 ciphers. Before when the 56bit ciphers were enabled the sorting was
2958 The syntax for the cipher sorting has been extended to support sorting by
2959 cipher-strength (using the strength_bits hard coded in the tables).
2960 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
2962 Fix a bug in the cipher-command parser: when supplying a cipher command
2963 string with an "undefined" symbol (neither command nor alphanumeric
2964 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
2965 an error is flagged.
2967 Due to the strength-sorting extension, the code of the
2968 ssl_create_cipher_list() function was completely rearranged. I hope that
2969 the readability was also increased :-)
2970 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
2972 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
2973 for the first serial number and places 2 in the serial number file. This
2974 avoids problems when the root CA is created with serial number zero and
2975 the first user certificate has the same issuer name and serial number
2979 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
2980 the new code. Add documentation for this stuff.
2983 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
2984 X509_*() to X509at_*() on the grounds that they don't handle X509
2985 structures and behave in an analagous way to the X509v3 functions:
2986 they shouldn't be called directly but wrapper functions should be used
2989 So we also now have some wrapper functions that call the X509at functions
2990 when passed certificate requests. (TO DO: similar things can be done with
2991 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
2992 things. Some of these need some d2i or i2d and print functionality
2993 because they handle more complex structures.)
2996 *) Add missing #ifndefs that caused missing symbols when building libssl
2997 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
2998 NO_RSA in ssl/s2*.c.
2999 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
3001 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
3002 has a return value which indicates the quality of the random data
3003 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
3004 error queue. New function RAND_pseudo_bytes() generates output that is
3005 guaranteed to be unique but not unpredictable. RAND_add is like
3006 RAND_seed, but takes an extra argument for an entropy estimate
3007 (RAND_seed always assumes full entropy).
3010 *) Do more iterations of Rabin-Miller probable prime test (specifically,
3011 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
3012 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
3013 in crypto/bn/bn_prime.c for the complete table). This guarantees a
3014 false-positive rate of at most 2^-80 for random input.
3017 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
3020 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
3021 in the 0.9.5 release), this returns the chain
3022 from an X509_CTX structure with a dup of the stack and all
3023 the X509 reference counts upped: so the stack will exist
3024 after X509_CTX_cleanup() has been called. Modify pkcs12.c
3027 Also make SSL_SESSION_print() print out the verify return
3031 *) Add manpage for the pkcs12 command. Also change the default
3032 behaviour so MAC iteration counts are used unless the new
3033 -nomaciter option is used. This improves file security and
3034 only older versions of MSIE (4.0 for example) need it.
3037 *) Honor the no-xxx Configure options when creating .DEF files.
3040 *) Add PKCS#10 attributes to field table: challengePassword,
3041 unstructuredName and unstructuredAddress. These are taken from
3042 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
3043 international characters are used.
3045 More changes to X509_ATTRIBUTE code: allow the setting of types
3046 based on strings. Remove the 'loc' parameter when adding
3047 attributes because these will be a SET OF encoding which is sorted
3051 *) Initial changes to the 'req' utility to allow request generation
3052 automation. This will allow an application to just generate a template
3053 file containing all the field values and have req construct the
3056 Initial support for X509_ATTRIBUTE handling. Stacks of these are
3057 used all over the place including certificate requests and PKCS#7
3058 structures. They are currently handled manually where necessary with
3059 some primitive wrappers for PKCS#7. The new functions behave in a
3060 manner analogous to the X509 extension functions: they allow
3061 attributes to be looked up by NID and added.
3063 Later something similar to the X509V3 code would be desirable to
3064 automatically handle the encoding, decoding and printing of the
3065 more complex types. The string types like challengePassword can
3066 be handled by the string table functions.
3068 Also modified the multi byte string table handling. Now there is
3069 a 'global mask' which masks out certain types. The table itself
3070 can use the flag STABLE_NO_MASK to ignore the mask setting: this
3071 is useful when for example there is only one permissible type
3072 (as in countryName) and using the mask might result in no valid
3076 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
3077 SSL_get_peer_finished to allow applications to obtain the latest
3078 Finished messages sent to the peer or expected from the peer,
3079 respectively. (SSL_get_peer_finished is usually the Finished message
3080 actually received from the peer, otherwise the protocol will be aborted.)
3082 As the Finished message are message digests of the complete handshake
3083 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
3084 be used for external authentication procedures when the authentication
3085 provided by SSL/TLS is not desired or is not enough.
3088 *) Enhanced support for Alpha Linux is added. Now ./config checks if
3089 the host supports BWX extension and if Compaq C is present on the
3090 $PATH. Just exploiting of the BWX extension results in 20-30%
3091 performance kick for some algorithms, e.g. DES and RC4 to mention
3092 a couple. Compaq C in turn generates ~20% faster code for MD5 and
3096 *) Add support for MS "fast SGC". This is arguably a violation of the
3097 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
3098 weak crypto and after checking the certificate is SGC a second one
3099 with strong crypto. MS SGC stops the first handshake after receiving
3100 the server certificate message and sends a second client hello. Since
3101 a server will typically do all the time consuming operations before
3102 expecting any further messages from the client (server key exchange
3103 is the most expensive) there is little difference between the two.
3105 To get OpenSSL to support MS SGC we have to permit a second client
3106 hello message after we have sent server done. In addition we have to
3107 reset the MAC if we do get this second client hello.
3110 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
3111 if a DER encoded private key is RSA or DSA traditional format. Changed
3112 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
3113 format DER encoded private key. Newer code should use PKCS#8 format which
3114 has the key type encoded in the ASN1 structure. Added DER private key
3115 support to pkcs8 application.
3118 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
3119 ciphersuites has been selected (as required by the SSL 3/TLS 1
3120 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
3121 is set, we interpret this as a request to violate the specification
3122 (the worst that can happen is a handshake failure, and 'correct'
3123 behaviour would result in a handshake failure anyway).
3126 *) In SSL_CTX_add_session, take into account that there might be multiple
3127 SSL_SESSION structures with the same session ID (e.g. when two threads
3128 concurrently obtain them from an external cache).
3129 The internal cache can handle only one SSL_SESSION with a given ID,
3130 so if there's a conflict, we now throw out the old one to achieve
3134 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
3135 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
3136 some routines that use cipher OIDs: some ciphers do not have OIDs
3137 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
3141 *) Simplify the trust setting structure and code. Now we just have
3142 two sequences of OIDs for trusted and rejected settings. These will
3143 typically have values the same as the extended key usage extension
3144 and any application specific purposes.
3146 The trust checking code now has a default behaviour: it will just
3147 check for an object with the same NID as the passed id. Functions can
3148 be provided to override either the default behaviour or the behaviour
3149 for a given id. SSL client, server and email already have functions
3150 in place for compatibility: they check the NID and also return "trusted"
3151 if the certificate is self signed.
3154 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
3155 traditional format into an EVP_PKEY structure.
3158 *) Add a password callback function PEM_cb() which either prompts for
3159 a password if usr_data is NULL or otherwise assumes it is a null
3160 terminated password. Allow passwords to be passed on command line
3161 environment or config files in a few more utilities.
3164 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
3165 keys. Add some short names for PKCS#8 PBE algorithms and allow them
3166 to be specified on the command line for the pkcs8 and pkcs12 utilities.
3167 Update documentation.
3170 *) Support for ASN1 "NULL" type. This could be handled before by using
3171 ASN1_TYPE but there wasn't any function that would try to read a NULL
3172 and produce an error if it couldn't. For compatibility we also have
3173 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
3174 don't allocate anything because they don't need to.
3177 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
3179 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
3181 *) Rebuild of the memory allocation routines used by OpenSSL code and
3182 possibly others as well. The purpose is to make an interface that
3183 provide hooks so anyone can build a separate set of allocation and
3184 deallocation routines to be used by OpenSSL, for example memory
3185 pool implementations, or something else, which was previously hard
3186 since Malloc(), Realloc() and Free() were defined as macros having
3187 the values malloc, realloc and free, respectively (except for Win32
3188 compilations). The same is provided for memory debugging code.
3189 OpenSSL already comes with functionality to find memory leaks, but
3190 this gives people a chance to debug other memory problems.
3192 With these changes, a new set of functions and macros have appeared:
3194 CRYPTO_set_mem_debug_functions() [F]
3195 CRYPTO_get_mem_debug_functions() [F]
3196 CRYPTO_dbg_set_options() [F]
3197 CRYPTO_dbg_get_options() [F]
3198 CRYPTO_malloc_debug_init() [M]
3200 The memory debug functions are NULL by default, unless the library
3201 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
3202 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
3203 gives the standard debugging functions that come with OpenSSL) or
3204 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
3205 provided by the library user) must be used. When the standard
3206 debugging functions are used, CRYPTO_dbg_set_options can be used to
3207 request additional information:
3208 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
3209 the CRYPTO_MDEBUG_xxx macro when compiling the library.
3211 Also, things like CRYPTO_set_mem_functions will always give the
3212 expected result (the new set of functions is used for allocation
3213 and deallocation) at all times, regardless of platform and compiler
3216 To finish it up, some functions that were never use in any other
3217 way than through macros have a new API and new semantic:
3220 CRYPTO_dbg_realloc()
3223 All macros of value have retained their old syntax.
3224 [Richard Levitte and Bodo Moeller]
3226 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
3227 ordering of SMIMECapabilities wasn't in "strength order" and there
3228 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
3232 *) Some ASN1 types with illegal zero length encoding (INTEGER,
3233 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
3234 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
3236 *) Merge in my S/MIME library for OpenSSL. This provides a simple
3237 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
3238 functionality to handle multipart/signed properly) and a utility
3239 called 'smime' to call all this stuff. This is based on code I
3240 originally wrote for Celo who have kindly allowed it to be
3241 included in OpenSSL.
3244 *) Add variants des_set_key_checked and des_set_key_unchecked of
3245 des_set_key (aka des_key_sched). Global variable des_check_key
3246 decides which of these is called by des_set_key; this way
3247 des_check_key behaves as it always did, but applications and
3248 the library itself, which was buggy for des_check_key == 1,
3249 have a cleaner way to pick the version they need.
3252 *) New function PKCS12_newpass() which changes the password of a
3256 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
3257 dynamic mix. In both cases the ids can be used as an index into the
3258 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
3259 functions so they accept a list of the field values and the
3260 application doesn't need to directly manipulate the X509_TRUST
3264 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
3268 *) Modify the way the V3 extension code looks up extensions. This now
3269 works in a similar way to the object code: we have some "standard"
3270 extensions in a static table which is searched with OBJ_bsearch()
3271 and the application can add dynamic ones if needed. The file
3272 crypto/x509v3/ext_dat.h now has the info: this file needs to be
3273 updated whenever a new extension is added to the core code and kept
3274 in ext_nid order. There is a simple program 'tabtest.c' which checks
3275 this. New extensions are not added too often so this file can readily
3276 be maintained manually.
3278 There are two big advantages in doing things this way. The extensions
3279 can be looked up immediately and no longer need to be "added" using
3280 X509V3_add_standard_extensions(): this function now does nothing.
3281 [Side note: I get *lots* of email saying the extension code doesn't
3282 work because people forget to call this function]
3283 Also no dynamic allocation is done unless new extensions are added:
3284 so if we don't add custom extensions there is no need to call
3285 X509V3_EXT_cleanup().
3288 *) Modify enc utility's salting as follows: make salting the default. Add a
3289 magic header, so unsalted files fail gracefully instead of just decrypting
3290 to garbage. This is because not salting is a big security hole, so people
3291 should be discouraged from doing it.
3294 *) Fixes and enhancements to the 'x509' utility. It allowed a message
3295 digest to be passed on the command line but it only used this
3296 parameter when signing a certificate. Modified so all relevant
3297 operations are affected by the digest parameter including the
3298 -fingerprint and -x509toreq options. Also -x509toreq choked if a
3299 DSA key was used because it didn't fix the digest.
3302 *) Initial certificate chain verify code. Currently tests the untrusted
3303 certificates for consistency with the verify purpose (which is set
3304 when the X509_STORE_CTX structure is set up) and checks the pathlength.
3306 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
3307 this is because it will reject chains with invalid extensions whereas
3308 every previous version of OpenSSL and SSLeay made no checks at all.
3310 Trust code: checks the root CA for the relevant trust settings. Trust
3311 settings have an initial value consistent with the verify purpose: e.g.
3312 if the verify purpose is for SSL client use it expects the CA to be
3313 trusted for SSL client use. However the default value can be changed to
3314 permit custom trust settings: one example of this would be to only trust
3315 certificates from a specific "secure" set of CAs.
3317 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
3318 which should be used for version portability: especially since the
3319 verify structure is likely to change more often now.
3321 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
3322 to set them. If not set then assume SSL clients will verify SSL servers
3325 Two new options to the verify program: -untrusted allows a set of
3326 untrusted certificates to be passed in and -purpose which sets the
3327 intended purpose of the certificate. If a purpose is set then the
3328 new chain verify code is used to check extension consistency.
3331 *) Support for the authority information access extension.
3334 *) Modify RSA and DSA PEM read routines to transparently handle
3335 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
3336 public keys in a format compatible with certificate
3337 SubjectPublicKeyInfo structures. Unfortunately there were already
3338 functions called *_PublicKey_* which used various odd formats so
3339 these are retained for compatibility: however the DSA variants were
3340 never in a public release so they have been deleted. Changed dsa/rsa
3341 utilities to handle the new format: note no releases ever handled public
3342 keys so we should be OK.
3344 The primary motivation for this change is to avoid the same fiasco
3345 that dogs private keys: there are several incompatible private key
3346 formats some of which are standard and some OpenSSL specific and
3347 require various evil hacks to allow partial transparent handling and
3348 even then it doesn't work with DER formats. Given the option anything
3349 other than PKCS#8 should be dumped: but the other formats have to
3350 stay in the name of compatibility.
3352 With public keys and the benefit of hindsight one standard format
3353 is used which works with EVP_PKEY, RSA or DSA structures: though
3354 it clearly returns an error if you try to read the wrong kind of key.
3356 Added a -pubkey option to the 'x509' utility to output the public key.
3357 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
3358 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
3359 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
3360 that do the same as the EVP_PKEY_assign_*() except they up the
3361 reference count of the added key (they don't "swallow" the
3365 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
3366 CRLs would fail if the file contained no certificates or no CRLs:
3367 added a new function to read in both types and return the number
3368 read: this means that if none are read it will be an error. The
3369 DER versions of the certificate and CRL reader would always fail
3370 because it isn't possible to mix certificates and CRLs in DER format
3371 without choking one or the other routine. Changed this to just read
3372 a certificate: this is the best we can do. Also modified the code
3373 in apps/verify.c to take notice of return codes: it was previously
3374 attempting to read in certificates from NULL pointers and ignoring
3375 any errors: this is one reason why the cert and CRL reader seemed
3376 to work. It doesn't check return codes from the default certificate
3377 routines: these may well fail if the certificates aren't installed.
3380 *) Code to support otherName option in GeneralName.
3383 *) First update to verify code. Change the verify utility
3384 so it warns if it is passed a self signed certificate:
3385 for consistency with the normal behaviour. X509_verify
3386 has been modified to it will now verify a self signed
3387 certificate if *exactly* the same certificate appears
3388 in the store: it was previously impossible to trust a
3389 single self signed certificate. This means that:
3390 openssl verify ss.pem
3391 now gives a warning about a self signed certificate but
3392 openssl verify -CAfile ss.pem ss.pem
3396 *) For servers, store verify_result in SSL_SESSION data structure
3397 (and add it to external session representation).
3398 This is needed when client certificate verifications fails,
3399 but an application-provided verification callback (set by
3400 SSL_CTX_set_cert_verify_callback) allows accepting the session
3401 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
3402 but returns 1): When the session is reused, we have to set
3403 ssl->verify_result to the appropriate error code to avoid
3405 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
3407 *) Fix a bug in the new PKCS#7 code: it didn't consider the
3408 case in PKCS7_dataInit() where the signed PKCS7 structure
3409 didn't contain any existing data because it was being created.
3410 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
3412 *) Add a salt to the key derivation routines in enc.c. This
3413 forms the first 8 bytes of the encrypted file. Also add a
3414 -S option to allow a salt to be input on the command line.
3417 *) New function X509_cmp(). Oddly enough there wasn't a function
3418 to compare two certificates. We do this by working out the SHA1
3419 hash and comparing that. X509_cmp() will be needed by the trust
3423 *) SSL_get1_session() is like SSL_get_session(), but increments
3424 the reference count in the SSL_SESSION returned.
3425 [Geoff Thorpe <geoff@eu.c2.net>]
3427 *) Fix for 'req': it was adding a null to request attributes.
3428 Also change the X509_LOOKUP and X509_INFO code to handle
3429 certificate auxiliary information.
3432 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
3436 *) Add the possibility to add extra information to the memory leak
3437 detecting output, to form tracebacks, showing from where each
3438 allocation was originated: CRYPTO_push_info("constant string") adds
3439 the string plus current file name and line number to a per-thread
3440 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
3441 is like calling CYRPTO_pop_info() until the stack is empty.
3442 Also updated memory leak detection code to be multi-thread-safe.
3445 *) Add options -text and -noout to pkcs7 utility and delete the
3446 encryption options which never did anything. Update docs.
3449 *) Add options to some of the utilities to allow the pass phrase
3450 to be included on either the command line (not recommended on
3451 OSes like Unix) or read from the environment. Update the
3452 manpages and fix a few bugs.
3455 *) Add a few manpages for some of the openssl commands.
3458 *) Fix the -revoke option in ca. It was freeing up memory twice,
3459 leaking and not finding already revoked certificates.
3462 *) Extensive changes to support certificate auxiliary information.
3463 This involves the use of X509_CERT_AUX structure and X509_AUX
3464 functions. An X509_AUX function such as PEM_read_X509_AUX()
3465 can still read in a certificate file in the usual way but it
3466 will also read in any additional "auxiliary information". By
3467 doing things this way a fair degree of compatibility can be
3468 retained: existing certificates can have this information added
3469 using the new 'x509' options.
3471 Current auxiliary information includes an "alias" and some trust
3472 settings. The trust settings will ultimately be used in enhanced
3473 certificate chain verification routines: currently a certificate
3474 can only be trusted if it is self signed and then it is trusted
3478 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
3479 The problem was that one of the replacement routines had not been working
3480 since SSLeay releases. For now the offending routine has been replaced
3481 with non-optimised assembler. Even so, this now gives around 95%
3482 performance improvement for 1024 bit RSA signs.
3485 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
3486 handling. Most clients have the effective key size in bits equal to
3487 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
3488 A few however don't do this and instead use the size of the decrypted key
3489 to determine the RC2 key length and the AlgorithmIdentifier to determine
3490 the effective key length. In this case the effective key length can still
3491 be 40 bits but the key length can be 168 bits for example. This is fixed
3492 by manually forcing an RC2 key into the EVP_PKEY structure because the
3493 EVP code can't currently handle unusual RC2 key sizes: it always assumes
3494 the key length and effective key length are equal.
3497 *) Add a bunch of functions that should simplify the creation of
3498 X509_NAME structures. Now you should be able to do:
3499 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
3500 and have it automatically work out the correct field type and fill in
3501 the structures. The more adventurous can try:
3502 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
3503 and it will (hopefully) work out the correct multibyte encoding.
3506 *) Change the 'req' utility to use the new field handling and multibyte
3507 copy routines. Before the DN field creation was handled in an ad hoc
3508 way in req, ca, and x509 which was rather broken and didn't support
3509 BMPStrings or UTF8Strings. Since some software doesn't implement
3510 BMPStrings or UTF8Strings yet, they can be enabled using the config file
3511 using the dirstring_type option. See the new comment in the default
3512 openssl.cnf for more info.
3515 *) Make crypto/rand/md_rand.c more robust:
3516 - Assure unique random numbers after fork().
3517 - Make sure that concurrent threads access the global counter and
3518 md serializably so that we never lose entropy in them
3519 or use exactly the same state in multiple threads.
3520 Access to the large state is not always serializable because
3521 the additional locking could be a performance killer, and
3522 md should be large enough anyway.
3525 *) New file apps/app_rand.c with commonly needed functionality
3526 for handling the random seed file.
3528 Use the random seed file in some applications that previously did not:
3530 dsaparam -genkey (which also ignored its '-rand' option),
3533 x509 (when signing).
3534 Except on systems with /dev/urandom, it is crucial to have a random
3535 seed file at least for key creation, DSA signing, and for DH exchanges;
3536 for RSA signatures we could do without one.
3538 gendh and gendsa (unlike genrsa) used to read only the first byte
3539 of each file listed in the '-rand' option. The function as previously
3540 found in genrsa is now in app_rand.c and is used by all programs
3541 that support '-rand'.
3544 *) In RAND_write_file, use mode 0600 for creating files;
3545 don't just chmod when it may be too late.
3548 *) Report an error from X509_STORE_load_locations
3549 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
3552 *) New function ASN1_mbstring_copy() this copies a string in either
3553 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
3554 into an ASN1_STRING type. A mask of permissible types is passed
3555 and it chooses the "minimal" type to use or an error if not type
3559 *) Add function equivalents to the various macros in asn1.h. The old
3560 macros are retained with an M_ prefix. Code inside the library can
3561 use the M_ macros. External code (including the openssl utility)
3562 should *NOT* in order to be "shared library friendly".
3565 *) Add various functions that can check a certificate's extensions
3566 to see if it usable for various purposes such as SSL client,
3567 server or S/MIME and CAs of these types. This is currently
3568 VERY EXPERIMENTAL but will ultimately be used for certificate chain
3569 verification. Also added a -purpose flag to x509 utility to
3570 print out all the purposes.
3573 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
3577 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
3578 for, obtain and decode and extension and obtain its critical flag.
3579 This allows all the necessary extension code to be handled in a
3580 single function call.
3583 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
3584 platforms. See crypto/rc4/rc4_enc.c for further details.
3587 *) New -noout option to asn1parse. This causes no output to be produced
3588 its main use is when combined with -strparse and -out to extract data
3589 from a file (which may not be in ASN.1 format).
3592 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
3593 when producing the local key id.
3594 [Richard Levitte <levitte@stacken.kth.se>]
3596 *) New option -dhparam in s_server. This allows a DH parameter file to be
3597 stated explicitly. If it is not stated then it tries the first server
3598 certificate file. The previous behaviour hard coded the filename
3602 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
3603 a public key to be input or output. For example:
3604 openssl rsa -in key.pem -pubout -out pubkey.pem
3605 Also added necessary DSA public key functions to handle this.
3608 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
3609 in the message. This was handled by allowing
3610 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
3611 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
3613 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
3614 to the end of the strings whereas this didn't. This would cause problems
3615 if strings read with d2i_ASN1_bytes() were later modified.
3616 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
3618 *) Fix for base64 decode bug. When a base64 bio reads only one line of
3619 data and it contains EOF it will end up returning an error. This is
3620 caused by input 46 bytes long. The cause is due to the way base64
3621 BIOs find the start of base64 encoded data. They do this by trying a
3622 trial decode on each line until they find one that works. When they
3623 do a flag is set and it starts again knowing it can pass all the
3624 data directly through the decoder. Unfortunately it doesn't reset
3625 the context it uses. This means that if EOF is reached an attempt
3626 is made to pass two EOFs through the context and this causes the
3627 resulting error. This can also cause other problems as well. As is
3628 usual with these problems it takes *ages* to find and the fix is
3629 trivial: move one line.
3630 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
3632 *) Ugly workaround to get s_client and s_server working under Windows. The
3633 old code wouldn't work because it needed to select() on sockets and the
3634 tty (for keypresses and to see if data could be written). Win32 only
3635 supports select() on sockets so we select() with a 1s timeout on the
3636 sockets and then see if any characters are waiting to be read, if none
3637 are present then we retry, we also assume we can always write data to
3638 the tty. This isn't nice because the code then blocks until we've
3639 received a complete line of data and it is effectively polling the
3640 keyboard at 1s intervals: however it's quite a bit better than not
3641 working at all :-) A dedicated Windows application might handle this
3642 with an event loop for example.
3645 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
3646 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
3647 will be called when RSA_sign() and RSA_verify() are used. This is useful
3648 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
3649 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
3650 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
3651 This necessitated the support of an extra signature type NID_md5_sha1
3652 for SSL signatures and modifications to the SSL library to use it instead
3653 of calling RSA_public_decrypt() and RSA_private_encrypt().
3656 *) Add new -verify -CAfile and -CApath options to the crl program, these
3657 will lookup a CRL issuers certificate and verify the signature in a
3658 similar way to the verify program. Tidy up the crl program so it
3659 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
3660 less strict. It will now permit CRL extensions even if it is not
3661 a V2 CRL: this will allow it to tolerate some broken CRLs.
3664 *) Initialize all non-automatic variables each time one of the openssl
3665 sub-programs is started (this is necessary as they may be started
3666 multiple times from the "OpenSSL>" prompt).
3667 [Lennart Bang, Bodo Moeller]
3669 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
3670 removing all other RSA functionality (this is what NO_RSA does). This
3671 is so (for example) those in the US can disable those operations covered
3672 by the RSA patent while allowing storage and parsing of RSA keys and RSA
3676 *) Non-copying interface to BIO pairs.
3677 (still largely untested)
3680 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
3681 ASCII string. This was handled independently in various places before.
3684 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
3685 UTF8 strings a character at a time.
3688 *) Use client_version from client hello to select the protocol
3689 (s23_srvr.c) and for RSA client key exchange verification
3690 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
3693 *) Add various utility functions to handle SPKACs, these were previously
3694 handled by poking round in the structure internals. Added new function
3695 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
3696 print, verify and generate SPKACs. Based on an original idea from
3697 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
3700 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
3703 *) Allow the config file extension section to be overwritten on the
3704 command line. Based on an original idea from Massimiliano Pala
3705 <madwolf@comune.modena.it>. The new option is called -extensions
3706 and can be applied to ca, req and x509. Also -reqexts to override
3707 the request extensions in req and -crlexts to override the crl extensions
3711 *) Add new feature to the SPKAC handling in ca. Now you can include
3712 the same field multiple times by preceding it by "XXXX." for example:
3715 this is the same syntax as used in the req config file.
3718 *) Allow certificate extensions to be added to certificate requests. These
3719 are specified in a 'req_extensions' option of the req section of the
3720 config file. They can be printed out with the -text option to req but
3721 are otherwise ignored at present.
3724 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
3725 data read consists of only the final block it would not decrypted because
3726 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
3727 A misplaced 'break' also meant the decrypted final block might not be
3728 copied until the next read.
3731 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
3732 a few extra parameters to the DH structure: these will be useful if
3733 for example we want the value of 'q' or implement X9.42 DH.
3736 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
3737 provides hooks that allow the default DSA functions or functions on a
3738 "per key" basis to be replaced. This allows hardware acceleration and
3739 hardware key storage to be handled without major modification to the
3740 library. Also added low level modexp hooks and CRYPTO_EX structure and
3741 associated functions.
3744 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
3745 as "read only": it can't be written to and the buffer it points to will
3746 not be freed. Reading from a read only BIO is much more efficient than
3747 a normal memory BIO. This was added because there are several times when
3748 an area of memory needs to be read from a BIO. The previous method was
3749 to create a memory BIO and write the data to it, this results in two
3750 copies of the data and an O(n^2) reading algorithm. There is a new
3751 function BIO_new_mem_buf() which creates a read only memory BIO from
3752 an area of memory. Also modified the PKCS#7 routines to use read only
3756 *) Bugfix: ssl23_get_client_hello did not work properly when called in
3757 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
3758 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
3759 but a retry condition occured while trying to read the rest.
3762 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
3763 NID_pkcs7_encrypted by default: this was wrong since this should almost
3764 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
3765 the encrypted data type: this is a more sensible place to put it and it
3766 allows the PKCS#12 code to be tidied up that duplicated this
3770 *) Changed obj_dat.pl script so it takes its input and output files on
3771 the command line. This should avoid shell escape redirection problems
3775 *) Initial support for certificate extension requests, these are included
3776 in things like Xenroll certificate requests. Included functions to allow
3777 extensions to be obtained and added.
3780 *) -crlf option to s_client and s_server for sending newlines as
3781 CRLF (as required by many protocols).
3784 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
3786 *) Install libRSAglue.a when OpenSSL is built with RSAref.
3787 [Ralf S. Engelschall]
3789 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
3790 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
3792 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
3796 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
3797 DH parameters/keys (q is lost during that conversion, but the resulting
3798 DH parameters contain its length).
3800 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
3801 much faster than DH_generate_parameters (which creates parameters
3802 where p = 2*q + 1), and also the smaller q makes DH computations
3803 much more efficient (160-bit exponentiation instead of 1024-bit
3804 exponentiation); so this provides a convenient way to support DHE
3805 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
3806 utter importance to use
3807 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
3809 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
3810 when such DH parameters are used, because otherwise small subgroup
3811 attacks may become possible!
3814 *) Avoid memory leak in i2d_DHparams.
3817 *) Allow the -k option to be used more than once in the enc program:
3818 this allows the same encrypted message to be read by multiple recipients.
3821 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
3822 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
3823 it will always use the numerical form of the OID, even if it has a short
3827 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
3828 method only got called if p,q,dmp1,dmq1,iqmp components were present,
3829 otherwise bn_mod_exp was called. In the case of hardware keys for example
3830 no private key components need be present and it might store extra data
3831 in the RSA structure, which cannot be accessed from bn_mod_exp.
3832 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
3833 private key operations.
3836 *) Added support for SPARC Linux.
3839 *) pem_password_cb function type incompatibly changed from
3840 typedef int pem_password_cb(char *buf, int size, int rwflag);
3842 ....(char *buf, int size, int rwflag, void *userdata);
3843 so that applications can pass data to their callbacks:
3844 The PEM[_ASN1]_{read,write}... functions and macros now take an
3845 additional void * argument, which is just handed through whenever
3846 the password callback is called.
3847 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
3849 New function SSL_CTX_set_default_passwd_cb_userdata.
3851 Compatibility note: As many C implementations push function arguments
3852 onto the stack in reverse order, the new library version is likely to
3853 interoperate with programs that have been compiled with the old
3854 pem_password_cb definition (PEM_whatever takes some data that
3855 happens to be on the stack as its last argument, and the callback
3856 just ignores this garbage); but there is no guarantee whatsoever that
3859 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
3860 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
3861 problems not only on Windows, but also on some Unix platforms.
3862 To avoid problematic command lines, these definitions are now in an
3863 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
3864 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
3867 *) MIPS III/IV assembler module is reimplemented.
3870 *) More DES library cleanups: remove references to srand/rand and
3871 delete an unused file.
3874 *) Add support for the the free Netwide assembler (NASM) under Win32,
3875 since not many people have MASM (ml) and it can be hard to obtain.
3876 This is currently experimental but it seems to work OK and pass all
3877 the tests. Check out INSTALL.W32 for info.
3880 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
3881 without temporary keys kept an extra copy of the server key,
3882 and connections with temporary keys did not free everything in case
3886 *) New function RSA_check_key and new openssl rsa option -check
3887 for verifying the consistency of RSA keys.
3888 [Ulf Moeller, Bodo Moeller]
3890 *) Various changes to make Win32 compile work:
3891 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
3892 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
3893 comparison" warnings.
3894 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
3897 *) Add a debugging option to PKCS#5 v2 key generation function: when
3898 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
3899 derived keys are printed to stderr.
3902 *) Copy the flags in ASN1_STRING_dup().
3903 [Roman E. Pavlov <pre@mo.msk.ru>]
3905 *) The x509 application mishandled signing requests containing DSA
3906 keys when the signing key was also DSA and the parameters didn't match.
3908 It was supposed to omit the parameters when they matched the signing key:
3909 the verifying software was then supposed to automatically use the CA's
3910 parameters if they were absent from the end user certificate.
3912 Omitting parameters is no longer recommended. The test was also
3913 the wrong way round! This was probably due to unusual behaviour in
3914 EVP_cmp_parameters() which returns 1 if the parameters match.
3915 This meant that parameters were omitted when they *didn't* match and
3916 the certificate was useless. Certificates signed with 'ca' didn't have
3918 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
3920 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
3921 The interface is as follows:
3922 Applications can use
3923 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
3924 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
3925 "off" is now the default.
3926 The library internally uses
3927 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
3928 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
3929 to disable memory-checking temporarily.
3931 Some inconsistent states that previously were possible (and were
3932 even the default) are now avoided.
3934 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
3935 with each memory chunk allocated; this is occasionally more helpful
3936 than just having a counter.
3938 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
3940 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
3944 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
3945 which largely parallels "options", but is for changing API behaviour,
3946 whereas "options" are about protocol behaviour.
3947 Initial "mode" flags are:
3949 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
3950 a single record has been written.
3951 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
3952 retries use the same buffer location.
3953 (But all of the contents must be
3957 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
3960 *) Fix problems with no-hmac etc.
3961 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
3963 *) New functions RSA_get_default_method(), RSA_set_method() and
3964 RSA_get_method(). These allows replacement of RSA_METHODs without having
3965 to mess around with the internals of an RSA structure.
3968 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
3969 Also really enable memory leak checks in openssl.c and in some
3971 [Chad C. Mulligan, Bodo Moeller]
3973 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
3974 up the length of negative integers. This has now been simplified to just
3975 store the length when it is first determined and use it later, rather
3976 than trying to keep track of where data is copied and updating it to
3978 [Steve Henson, reported by Brien Wheeler
3979 <bwheeler@authentica-security.com>]
3981 *) Add a new function PKCS7_signatureVerify. This allows the verification
3982 of a PKCS#7 signature but with the signing certificate passed to the
3983 function itself. This contrasts with PKCS7_dataVerify which assumes the
3984 certificate is present in the PKCS#7 structure. This isn't always the
3985 case: certificates can be omitted from a PKCS#7 structure and be
3986 distributed by "out of band" means (such as a certificate database).
3989 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
3990 function prototypes in pem.h, also change util/mkdef.pl to add the
3991 necessary function names.
3994 *) mk1mf.pl (used by Windows builds) did not properly read the
3995 options set by Configure in the top level Makefile, and Configure
3996 was not even able to write more than one option correctly.
3997 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
4000 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
4001 file to be loaded from a BIO or FILE pointer. The BIO version will
4002 for example allow memory BIOs to contain config info.
4005 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
4006 Whoever hopes to achieve shared-library compatibility across versions
4007 must use this, not the compile-time macro.
4008 (Exercise 0.9.4: Which is the minimum library version required by
4010 Note: All this applies only to multi-threaded programs, others don't
4014 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
4015 through a BIO pair triggered the default case, i.e.
4016 SSLerr(...,SSL_R_UNKNOWN_STATE).
4019 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
4020 can use the SSL library even if none of the specific BIOs is
4024 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
4025 for the encoded length.
4026 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
4028 *) Add initial documentation of the X509V3 functions.
4031 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
4032 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
4033 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
4034 secure PKCS#8 private key format with a high iteration count.
4037 *) Fix determination of Perl interpreter: A perl or perl5
4038 _directory_ in $PATH was also accepted as the interpreter.
4039 [Ralf S. Engelschall]
4041 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
4042 wrong with it but it was very old and did things like calling
4043 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
4047 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
4048 to use the new extension code.
4051 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
4052 with macros. This should make it easier to change their form, add extra
4053 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
4057 *) Add to configuration table a new entry that can specify an alternative
4058 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
4059 according to Mark Crispin <MRC@Panda.COM>.
4063 *) DES CBC did not update the IV. Weird.
4066 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
4067 Changing the behaviour of the former might break existing programs --
4068 where IV updating is needed, des_ncbc_encrypt can be used.
4071 *) When bntest is run from "make test" it drives bc to check its
4072 calculations, as well as internally checking them. If an internal check
4073 fails, it needs to cause bc to give a non-zero result or make test carries
4074 on without noticing the failure. Fixed.
4077 *) DES library cleanups.
4080 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
4081 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
4082 ciphers. NOTE: although the key derivation function has been verified
4083 against some published test vectors it has not been extensively tested
4084 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
4088 *) Instead of "mkdir -p", which is not fully portable, use new
4089 Perl script "util/mkdir-p.pl".
4092 *) Rewrite the way password based encryption (PBE) is handled. It used to
4093 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
4094 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
4095 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
4096 the 'parameter' field of the AlgorithmIdentifier is passed to the
4097 underlying key generation function so it must do its own ASN1 parsing.
4098 This has also changed the EVP_PBE_CipherInit() function which now has a
4099 'parameter' argument instead of literal salt and iteration count values
4100 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
4103 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
4104 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
4105 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
4106 KEY" because this clashed with PKCS#8 unencrypted string. Since this
4107 value was just used as a "magic string" and not used directly its
4108 value doesn't matter.
4111 *) Introduce some semblance of const correctness to BN. Shame C doesn't
4115 *) "linux-sparc64" configuration (ultrapenguin).
4116 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
4117 "linux-sparc" configuration.
4118 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
4120 *) config now generates no-xxx options for missing ciphers.
4123 *) Support the EBCDIC character set (work in progress).
4124 File ebcdic.c not yet included because it has a different license.
4125 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
4127 *) Support BS2000/OSD-POSIX.
4128 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
4130 *) Make callbacks for key generation use void * instead of char *.
4133 *) Make S/MIME samples compile (not yet tested).
4136 *) Additional typesafe stacks.
4139 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
4143 Changes between 0.9.3 and 0.9.3a [29 May 1999]
4145 *) New configuration variant "sco5-gcc".
4147 *) Updated some demos.
4148 [Sean O Riordain, Wade Scholine]
4150 *) Add missing BIO_free at exit of pkcs12 application.
4153 *) Fix memory leak in conf.c.
4156 *) Updates for Win32 to assembler version of MD5.
4159 *) Set #! path to perl in apps/der_chop to where we found it
4160 instead of using a fixed path.
4163 *) SHA library changes for irix64-mips4-cc.
4166 *) Improvements for VMS support.
4170 Changes between 0.9.2b and 0.9.3 [24 May 1999]
4172 *) Bignum library bug fix. IRIX 6 passes "make test" now!
4173 This also avoids the problems with SC4.2 and unpatched SC5.
4174 [Andy Polyakov <appro@fy.chalmers.se>]
4176 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
4177 These are required because of the typesafe stack would otherwise break
4178 existing code. If old code used a structure member which used to be STACK
4179 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
4180 sk_num or sk_value it would produce an error because the num, data members
4181 are not present in STACK_OF. Now it just produces a warning. sk_set
4182 replaces the old method of assigning a value to sk_value
4183 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
4184 that does this will no longer work (and should use sk_set instead) but
4185 this could be regarded as a "questionable" behaviour anyway.
4188 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
4189 correctly handle encrypted S/MIME data.
4192 *) Change type of various DES function arguments from des_cblock
4193 (which means, in function argument declarations, pointer to char)
4194 to des_cblock * (meaning pointer to array with 8 char elements),
4195 which allows the compiler to do more typechecking; it was like
4196 that back in SSLeay, but with lots of ugly casts.
4198 Introduce new type const_des_cblock.
4201 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
4202 problems: find RecipientInfo structure that matches recipient certificate
4203 and initialise the ASN1 structures properly based on passed cipher.
4206 *) Belatedly make the BN tests actually check the results.
4209 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
4210 to and from BNs: it was completely broken. New compilation option
4211 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
4212 key elements as negative integers.
4215 *) Reorganize and speed up MD5.
4216 [Andy Polyakov <appro@fy.chalmers.se>]
4219 [Richard Levitte <richard@levitte.org>]
4221 *) New option -out to asn1parse to allow the parsed structure to be
4222 output to a file. This is most useful when combined with the -strparse
4223 option to examine the output of things like OCTET STRINGS.
4226 *) Make SSL library a little more fool-proof by not requiring any longer
4227 that SSL_set_{accept,connect}_state be called before
4228 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
4229 in many applications because usually everything *appeared* to work as
4230 intended anyway -- now it really works as intended).
4233 *) Move openssl.cnf out of lib/.
4236 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
4237 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
4238 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
4239 [Ralf S. Engelschall]
4241 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
4242 handle PKCS#7 enveloped data properly.
4243 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
4245 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
4246 copying pointers. The cert_st handling is changed by this in
4247 various ways (and thus what used to be known as ctx->default_cert
4248 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
4249 any longer when s->cert does not give us what we need).
4250 ssl_cert_instantiate becomes obsolete by this change.
4251 As soon as we've got the new code right (possibly it already is?),
4252 we have solved a couple of bugs of the earlier code where s->cert
4253 was used as if it could not have been shared with other SSL structures.
4255 Note that using the SSL API in certain dirty ways now will result
4256 in different behaviour than observed with earlier library versions:
4257 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
4258 does not influence s as it used to.
4260 In order to clean up things more thoroughly, inside SSL_SESSION
4261 we don't use CERT any longer, but a new structure SESS_CERT
4262 that holds per-session data (if available); currently, this is
4263 the peer's certificate chain and, for clients, the server's certificate
4264 and temporary key. CERT holds only those values that can have
4265 meaningful defaults in an SSL_CTX.
4268 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
4269 from the internal representation. Various PKCS#7 fixes: remove some
4270 evil casts and set the enc_dig_alg field properly based on the signing
4274 *) Allow PKCS#12 password to be set from the command line or the
4275 environment. Let 'ca' get its config file name from the environment
4276 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
4280 *) Allow certificate policies extension to use an IA5STRING for the
4281 organization field. This is contrary to the PKIX definition but
4282 VeriSign uses it and IE5 only recognises this form. Document 'x509'
4286 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
4287 without disallowing inline assembler and the like for non-pedantic builds.
4290 *) Support Borland C++ builder.
4291 [Janez Jere <jj@void.si>, modified by Ulf Möller]
4296 *) SHA-1 cleanups and performance enhancements.
4297 [Andy Polyakov <appro@fy.chalmers.se>]
4299 *) Sparc v8plus assembler for the bignum library.
4300 [Andy Polyakov <appro@fy.chalmers.se>]
4302 *) Accept any -xxx and +xxx compiler options in Configure.
4305 *) Update HPUX configuration.
4308 *) Add missing sk_<type>_unshift() function to safestack.h
4309 [Ralf S. Engelschall]
4311 *) New function SSL_CTX_use_certificate_chain_file that sets the
4312 "extra_cert"s in addition to the certificate. (This makes sense
4313 only for "PEM" format files, as chains as a whole are not
4317 *) Support verify_depth from the SSL API.
4318 x509_vfy.c had what can be considered an off-by-one-error:
4319 Its depth (which was not part of the external interface)
4320 was actually counting the number of certificates in a chain;
4321 now it really counts the depth.
4324 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
4325 instead of X509err, which often resulted in confusing error
4326 messages since the error codes are not globally unique
4327 (e.g. an alleged error in ssl3_accept when a certificate
4328 didn't match the private key).
4330 *) New function SSL_CTX_set_session_id_context that allows to set a default
4331 value (so that you don't need SSL_set_session_id_context for each
4332 connection using the SSL_CTX).
4335 *) OAEP decoding bug fix.
4338 *) Support INSTALL_PREFIX for package builders, as proposed by
4342 *) New Configure options "threads" and "no-threads". For systems
4343 where the proper compiler options are known (currently Solaris
4344 and Linux), "threads" is the default.
4347 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
4350 *) Install various scripts to $(OPENSSLDIR)/misc, not to
4351 $(INSTALLTOP)/bin -- they shouldn't clutter directories
4352 such as /usr/local/bin.
4355 *) "make linux-shared" to build shared libraries.
4356 [Niels Poppe <niels@netbox.org>]
4358 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
4361 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
4362 extension adding in x509 utility.
4365 *) Remove NOPROTO sections and error code comments.
4368 *) Partial rewrite of the DEF file generator to now parse the ANSI
4372 *) New Configure options --prefix=DIR and --openssldir=DIR.
4375 *) Complete rewrite of the error code script(s). It is all now handled
4376 by one script at the top level which handles error code gathering,
4377 header rewriting and C source file generation. It should be much better
4378 than the old method: it now uses a modified version of Ulf's parser to
4379 read the ANSI prototypes in all header files (thus the old K&R definitions
4380 aren't needed for error creation any more) and do a better job of
4381 translating function codes into names. The old 'ASN1 error code imbedded
4382 in a comment' is no longer necessary and it doesn't use .err files which
4383 have now been deleted. Also the error code call doesn't have to appear all
4384 on one line (which resulted in some large lines...).
4387 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
4390 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
4391 0 (which usually indicates a closed connection), but continue reading.
4394 *) Fix some race conditions.
4397 *) Add support for CRL distribution points extension. Add Certificate
4398 Policies and CRL distribution points documentation.
4401 *) Move the autogenerated header file parts to crypto/opensslconf.h.
4404 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
4405 8 of keying material. Merlin has also confirmed interop with this fix
4406 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
4407 [Merlin Hughes <merlin@baltimore.ie>]
4409 *) Fix lots of warnings.
4410 [Richard Levitte <levitte@stacken.kth.se>]
4412 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
4413 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
4414 [Richard Levitte <levitte@stacken.kth.se>]
4416 *) Fix problems with sizeof(long) == 8.
4417 [Andy Polyakov <appro@fy.chalmers.se>]
4419 *) Change functions to ANSI C.
4422 *) Fix typos in error codes.
4423 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
4425 *) Remove defunct assembler files from Configure.
4428 *) SPARC v8 assembler BIGNUM implementation.
4429 [Andy Polyakov <appro@fy.chalmers.se>]
4431 *) Support for Certificate Policies extension: both print and set.
4432 Various additions to support the r2i method this uses.
4435 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
4436 return a const string when you are expecting an allocated buffer.
4439 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
4440 types DirectoryString and DisplayText.
4443 *) Add code to allow r2i extensions to access the configuration database,
4444 add an LHASH database driver and add several ctx helper functions.
4447 *) Fix an evil bug in bn_expand2() which caused various BN functions to
4448 fail when they extended the size of a BIGNUM.
4451 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
4452 support typesafe stack.
4455 *) Fix typo in SSL_[gs]et_options().
4456 [Nils Frostberg <nils@medcom.se>]
4458 *) Delete various functions and files that belonged to the (now obsolete)
4459 old X509V3 handling code.
4462 *) New Configure option "rsaref".
4465 *) Don't auto-generate pem.h.
4468 *) Introduce type-safe ASN.1 SETs.
4471 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
4472 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
4474 *) Introduce type-safe STACKs. This will almost certainly break lots of code
4475 that links with OpenSSL (well at least cause lots of warnings), but fear
4476 not: the conversion is trivial, and it eliminates loads of evil casts. A
4477 few STACKed things have been converted already. Feel free to convert more.
4478 In the fullness of time, I'll do away with the STACK type altogether.
4481 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
4482 specified in <certfile> by updating the entry in the index.txt file.
4483 This way one no longer has to edit the index.txt file manually for
4484 revoking a certificate. The -revoke option does the gory details now.
4485 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
4487 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
4488 `-text' option at all and this way the `-noout -text' combination was
4489 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
4490 [Ralf S. Engelschall]
4492 *) Make sure a corresponding plain text error message exists for the
4493 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
4494 verify callback function determined that a certificate was revoked.
4495 [Ralf S. Engelschall]
4497 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
4498 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
4499 all available cipers including rc5, which was forgotten until now.
4500 In order to let the testing shell script know which algorithms
4501 are available, a new (up to now undocumented) command
4502 "openssl list-cipher-commands" is used.
4505 *) Bugfix: s_client occasionally would sleep in select() when
4506 it should have checked SSL_pending() first.
4509 *) New functions DSA_do_sign and DSA_do_verify to provide access to
4510 the raw DSA values prior to ASN.1 encoding.
4513 *) Tweaks to Configure
4514 [Niels Poppe <niels@netbox.org>]
4516 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
4520 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
4523 *) New config option to avoid instructions that are illegal on the 80386.
4524 The default code is faster, but requires at least a 486.
4527 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
4528 SSL2_SERVER_VERSION (not used at all) macros, which are now the
4529 same as SSL2_VERSION anyway.
4532 *) New "-showcerts" option for s_client.
4535 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
4536 application. Various cleanups and fixes.
4539 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
4540 modify error routines to work internally. Add error codes and PBE init
4541 to library startup routines.
4544 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
4545 packing functions to asn1 and evp. Changed function names and error
4546 codes along the way.
4549 *) PKCS12 integration: and so it begins... First of several patches to
4550 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
4551 objects to objects.h
4554 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
4555 and display support for Thawte strong extranet extension.
4558 *) Add LinuxPPC support.
4559 [Jeff Dubrule <igor@pobox.org>]
4561 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
4562 bn_div_words in alpha.s.
4563 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
4565 *) Make sure the RSA OAEP test is skipped under -DRSAref because
4566 OAEP isn't supported when OpenSSL is built with RSAref.
4567 [Ulf Moeller <ulf@fitug.de>]
4569 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
4570 so they no longer are missing under -DNOPROTO.
4571 [Soren S. Jorvang <soren@t.dk>]
4574 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
4576 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
4577 doesn't work when the session is reused. Coming soon!
4580 *) Fix a security hole, that allows sessions to be reused in the wrong
4581 context thus bypassing client cert protection! All software that uses
4582 client certs and session caches in multiple contexts NEEDS PATCHING to
4583 allow session reuse! A fuller solution is in the works.
4584 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
4586 *) Some more source tree cleanups (removed obsolete files
4587 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
4588 permission on "config" script to be executable) and a fix for the INSTALL
4590 [Ulf Moeller <ulf@fitug.de>]
4592 *) Remove some legacy and erroneous uses of malloc, free instead of
4594 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
4596 *) Make rsa_oaep_test return non-zero on error.
4597 [Ulf Moeller <ulf@fitug.de>]
4599 *) Add support for native Solaris shared libraries. Configure
4600 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
4601 if someone would make that last step automatic.
4602 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
4604 *) ctx_size was not built with the right compiler during "make links". Fixed.
4607 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
4608 except NULL ciphers". This means the default cipher list will no longer
4609 enable NULL ciphers. They need to be specifically enabled e.g. with
4610 the string "DEFAULT:eNULL".
4613 *) Fix to RSA private encryption routines: if p < q then it would
4614 occasionally produce an invalid result. This will only happen with
4615 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
4618 *) Be less restrictive and allow also `perl util/perlpath.pl
4619 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
4620 because this way one can also use an interpreter named `perl5' (which is
4621 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
4622 installed as `perl').
4623 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
4625 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
4626 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
4628 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
4629 advapi32.lib to Win32 build and change the pem test comparision
4630 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
4631 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
4632 and crypto/des/ede_cbcm_enc.c.
4635 *) DES quad checksum was broken on big-endian architectures. Fixed.
4638 *) Comment out two functions in bio.h that aren't implemented. Fix up the
4639 Win32 test batch file so it (might) work again. The Win32 test batch file
4640 is horrible: I feel ill....
4643 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
4644 in e_os.h. Audit of header files to check ANSI and non ANSI
4645 sections: 10 functions were absent from non ANSI section and not exported
4646 from Windows DLLs. Fixed up libeay.num for new functions.
4649 *) Make `openssl version' output lines consistent.
4650 [Ralf S. Engelschall]
4652 *) Fix Win32 symbol export lists for BIO functions: Added
4653 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
4654 to ms/libeay{16,32}.def.
4655 [Ralf S. Engelschall]
4657 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
4658 fine under Unix and passes some trivial tests I've now added. But the
4659 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
4660 added to make sure no one expects that this stuff really works in the
4661 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
4662 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
4664 [Ralf S. Engelschall]
4666 *) Fix the generation of two part addresses in perl.
4667 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
4669 *) Add config entry for Linux on MIPS.
4670 [John Tobey <jtobey@channel1.com>]
4672 *) Make links whenever Configure is run, unless we are on Windoze.
4675 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
4676 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
4680 *) Add a useful kludge to allow package maintainers to specify compiler and
4681 other platforms details on the command line without having to patch the
4682 Configure script everytime: One now can use ``perl Configure
4683 <id>:<details>'', i.e. platform ids are allowed to have details appended
4684 to them (seperated by colons). This is treated as there would be a static
4685 pre-configured entry in Configure's %table under key <id> with value
4686 <details> and ``perl Configure <id>'' is called. So, when you want to
4687 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
4688 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
4689 now, which overrides the FreeBSD-elf entry on-the-fly.
4690 [Ralf S. Engelschall]
4692 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
4695 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
4696 on the `perl Configure ...' command line. This way one can compile
4697 OpenSSL libraries with Position Independent Code (PIC) which is needed
4698 for linking it into DSOs.
4699 [Ralf S. Engelschall]
4701 *) Remarkably, export ciphers were totally broken and no-one had noticed!
4705 *) Cleaned up the LICENSE document: The official contact for any license
4706 questions now is the OpenSSL core team under openssl-core@openssl.org.
4707 And add a paragraph about the dual-license situation to make sure people
4708 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
4709 to the OpenSSL toolkit.
4710 [Ralf S. Engelschall]
4712 *) General source tree makefile cleanups: Made `making xxx in yyy...'
4713 display consistent in the source tree and replaced `/bin/rm' by `rm'.
4714 Additonally cleaned up the `make links' target: Remove unnecessary
4715 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
4716 to speed processing and no longer clutter the display with confusing
4717 stuff. Instead only the actually done links are displayed.
4718 [Ralf S. Engelschall]
4720 *) Permit null encryption ciphersuites, used for authentication only. It used
4721 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
4722 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
4726 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
4727 signed attributes when verifying signatures (this would break them),
4728 the detached data encoding was wrong and public keys obtained using
4729 X509_get_pubkey() weren't freed.
4732 *) Add text documentation for the BUFFER functions. Also added a work around
4733 to a Win95 console bug. This was triggered by the password read stuff: the
4734 last character typed gets carried over to the next fread(). If you were
4735 generating a new cert request using 'req' for example then the last
4736 character of the passphrase would be CR which would then enter the first
4740 *) Added the new `Includes OpenSSL Cryptography Software' button as
4741 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
4742 button and can be used by applications based on OpenSSL to show the
4743 relationship to the OpenSSL project.
4744 [Ralf S. Engelschall]
4746 *) Remove confusing variables in function signatures in files
4747 ssl/ssl_lib.c and ssl/ssl.h.
4748 [Lennart Bong <lob@kulthea.stacken.kth.se>]
4750 *) Don't install bss_file.c under PREFIX/include/
4751 [Lennart Bong <lob@kulthea.stacken.kth.se>]
4753 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
4754 functions that return function pointers and has support for NT specific
4755 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
4756 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
4757 unsigned to signed types: this was killing the Win32 compile.
4760 *) Add new certificate file to stack functions,
4761 SSL_add_dir_cert_subjects_to_stack() and
4762 SSL_add_file_cert_subjects_to_stack(). These largely supplant
4763 SSL_load_client_CA_file(), and can be used to add multiple certs easily
4764 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
4765 This means that Apache-SSL and similar packages don't have to mess around
4766 to add as many CAs as they want to the preferred list.
4769 *) Experiment with doxygen documentation. Currently only partially applied to
4771 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
4772 openssl.doxy as the configuration file.
4775 *) Get rid of remaining C++-style comments which strict C compilers hate.
4776 [Ralf S. Engelschall, pointed out by Carlos Amengual]
4778 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
4779 compiled in by default: it has problems with large keys.
4782 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
4783 DH private keys and/or callback functions which directly correspond to
4784 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
4785 is needed for applications which have to configure certificates on a
4786 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
4788 For the RSA certificate situation is makes no difference, but
4789 for the DSA certificate situation this fixes the "no shared cipher"
4790 problem where the OpenSSL cipher selection procedure failed because the
4791 temporary keys were not overtaken from the context and the API provided
4792 no way to reconfigure them.
4793 The new functions now let applications reconfigure the stuff and they
4794 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
4795 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
4796 non-public-API function ssl_cert_instantiate() is used as a helper
4797 function and also to reduce code redundancy inside ssl_rsa.c.
4798 [Ralf S. Engelschall]
4800 *) Move s_server -dcert and -dkey options out of the undocumented feature
4801 area because they are useful for the DSA situation and should be
4802 recognized by the users.
4803 [Ralf S. Engelschall]
4805 *) Fix the cipher decision scheme for export ciphers: the export bits are
4806 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
4807 SSL_EXP_MASK. So, the original variable has to be used instead of the
4808 already masked variable.
4809 [Richard Levitte <levitte@stacken.kth.se>]
4811 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
4812 [Richard Levitte <levitte@stacken.kth.se>]
4814 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
4815 from `int' to `unsigned int' because it's a length and initialized by
4816 EVP_DigestFinal() which expects an `unsigned int *'.
4817 [Richard Levitte <levitte@stacken.kth.se>]
4819 *) Don't hard-code path to Perl interpreter on shebang line of Configure
4820 script. Instead use the usual Shell->Perl transition trick.
4821 [Ralf S. Engelschall]
4823 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
4824 (in addition to RSA certificates) to match the behaviour of `openssl dsa
4825 -noout -modulus' as it's already the case for `openssl rsa -noout
4826 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
4827 currently the public key is printed (a decision which was already done by
4828 `openssl dsa -modulus' in the past) which serves a similar purpose.
4829 Additionally the NO_RSA no longer completely removes the whole -modulus
4830 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
4832 [Ralf S. Engelschall]
4834 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
4835 BIO. See the source (crypto/evp/bio_ok.c) for more info.
4836 [Arne Ansper <arne@ats.cyber.ee>]
4838 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
4839 to be added. Now both 'req' and 'ca' can use new objects defined in the
4843 *) Add cool BIO that does syslog (or event log on NT).
4844 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
4846 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
4847 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
4848 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
4849 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
4852 *) Add preliminary config info for new extension code.
4855 *) Make RSA_NO_PADDING really use no padding.
4856 [Ulf Moeller <ulf@fitug.de>]
4858 *) Generate errors when private/public key check is done.
4861 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
4862 for some CRL extensions and new objects added.
4865 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
4866 key usage extension and fuller support for authority key id.
4869 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
4870 padding method for RSA, which is recommended for new applications in PKCS
4871 #1 v2.0 (RFC 2437, October 1998).
4872 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
4873 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
4874 against Bleichbacher's attack on RSA.
4875 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
4878 *) Updates to the new SSL compression code
4879 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
4881 *) Fix so that the version number in the master secret, when passed
4882 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
4883 (because the server will not accept higher), that the version number
4884 is 0x03,0x01, not 0x03,0x00
4885 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
4887 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
4888 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
4889 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
4892 *) Support for RAW extensions where an arbitrary extension can be
4893 created by including its DER encoding. See apps/openssl.cnf for
4897 *) Make sure latest Perl versions don't interpret some generated C array
4898 code as Perl array code in the crypto/err/err_genc.pl script.
4899 [Lars Weber <3weber@informatik.uni-hamburg.de>]
4901 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
4902 not many people have the assembler. Various Win32 compilation fixes and
4903 update to the INSTALL.W32 file with (hopefully) more accurate Win32
4907 *) Modify configure script 'Configure' to automatically create crypto/date.h
4908 file under Win32 and also build pem.h from pem.org. New script
4909 util/mkfiles.pl to create the MINFO file on environments that can't do a
4910 'make files': perl util/mkfiles.pl >MINFO should work.
4913 *) Major rework of DES function declarations, in the pursuit of correctness
4914 and purity. As a result, many evil casts evaporated, and some weirdness,
4915 too. You may find this causes warnings in your code. Zapping your evil
4916 casts will probably fix them. Mostly.
4919 *) Fix for a typo in asn1.h. Bug fix to object creation script
4920 obj_dat.pl. It considered a zero in an object definition to mean
4921 "end of object": none of the objects in objects.h have any zeros
4922 so it wasn't spotted.
4923 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
4925 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
4926 Masking (CBCM). In the absence of test vectors, the best I have been able
4927 to do is check that the decrypt undoes the encrypt, so far. Send me test
4928 vectors if you have them.
4931 *) Correct calculation of key length for export ciphers (too much space was
4932 allocated for null ciphers). This has not been tested!
4935 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
4936 message is now correct (it understands "crypto" and "ssl" on its
4937 command line). There is also now an "update" option. This will update
4938 the util/ssleay.num and util/libeay.num files with any new functions.
4940 perl util/mkdef.pl crypto ssl update
4941 it will update them.
4944 *) Overhauled the Perl interface (perl/*):
4945 - ported BN stuff to OpenSSL's different BN library
4946 - made the perl/ source tree CVS-aware
4947 - renamed the package from SSLeay to OpenSSL (the files still contain
4948 their history because I've copied them in the repository)
4949 - removed obsolete files (the test scripts will be replaced
4950 by better Test::Harness variants in the future)
4951 [Ralf S. Engelschall]
4953 *) First cut for a very conservative source tree cleanup:
4954 1. merge various obsolete readme texts into doc/ssleay.txt
4955 where we collect the old documents and readme texts.
4956 2. remove the first part of files where I'm already sure that we no
4957 longer need them because of three reasons: either they are just temporary
4958 files which were left by Eric or they are preserved original files where
4959 I've verified that the diff is also available in the CVS via "cvs diff
4960 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
4961 the crypto/md/ stuff).
4962 [Ralf S. Engelschall]
4964 *) More extension code. Incomplete support for subject and issuer alt
4965 name, issuer and authority key id. Change the i2v function parameters
4966 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
4967 what that's for :-) Fix to ASN1 macro which messed up
4968 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
4971 *) Preliminary support for ENUMERATED type. This is largely copied from the
4975 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
4976 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
4978 *) Make sure `make rehash' target really finds the `openssl' program.
4979 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
4981 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
4982 like to hear about it if this slows down other processors.
4985 *) Add CygWin32 platform information to Configure script.
4986 [Alan Batie <batie@aahz.jf.intel.com>]
4988 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
4989 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
4991 *) New program nseq to manipulate netscape certificate sequences
4994 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
4998 *) Fixes to BN code. Previously the default was to define BN_RECURSION
4999 but the BN code had some problems that would cause failures when
5000 doing certificate verification and some other functions.
5001 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
5003 *) Add ASN1 and PEM code to support netscape certificate sequences.
5006 *) Add ASN1 and PEM code to support netscape certificate sequences.
5009 *) Add several PKIX and private extended key usage OIDs.
5012 *) Modify the 'ca' program to handle the new extension code. Modify
5013 openssl.cnf for new extension format, add comments.
5016 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
5017 and add a sample to openssl.cnf so req -x509 now adds appropriate
5021 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
5022 error code, add initial support to X509_print() and x509 application.
5025 *) Takes a deep breath and start addding X509 V3 extension support code. Add
5026 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
5027 stuff is currently isolated and isn't even compiled yet.
5030 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
5031 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
5032 Removed the versions check from X509 routines when loading extensions:
5033 this allows certain broken certificates that don't set the version
5034 properly to be processed.
5037 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
5038 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
5039 can still be regenerated with "make depend".
5042 *) Spelling mistake in C version of CAST-128.
5043 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
5045 *) Changes to the error generation code. The perl script err-code.pl
5046 now reads in the old error codes and retains the old numbers, only
5047 adding new ones if necessary. It also only changes the .err files if new
5048 codes are added. The makefiles have been modified to only insert errors
5049 when needed (to avoid needlessly modifying header files). This is done
5050 by only inserting errors if the .err file is newer than the auto generated
5051 C file. To rebuild all the error codes from scratch (the old behaviour)
5052 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
5053 or delete all the .err files.
5056 *) CAST-128 was incorrectly implemented for short keys. The C version has
5057 been fixed, but is untested. The assembler versions are also fixed, but
5058 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
5059 to regenerate it if needed.
5060 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
5061 Hagino <itojun@kame.net>]
5063 *) File was opened incorrectly in randfile.c.
5064 [Ulf Möller <ulf@fitug.de>]
5066 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
5067 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
5068 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
5069 al: it's just almost always a UTCTime. Note this patch adds new error
5070 codes so do a "make errors" if there are problems.
5073 *) Correct Linux 1 recognition in config.
5074 [Ulf Möller <ulf@fitug.de>]
5076 *) Remove pointless MD5 hash when using DSA keys in ca.
5077 [Anonymous <nobody@replay.com>]
5079 *) Generate an error if given an empty string as a cert directory. Also
5080 generate an error if handed NULL (previously returned 0 to indicate an
5081 error, but didn't set one).
5082 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
5084 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
5087 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
5088 parameters. This was causing a warning which killed off the Win32 compile.
5091 *) Remove C++ style comments from crypto/bn/bn_local.h.
5092 [Neil Costigan <neil.costigan@celocom.com>]
5094 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
5095 based on a text string, looking up short and long names and finally
5096 "dot" format. The "dot" format stuff didn't work. Added new function
5097 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
5098 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
5099 OID is not part of the table.
5102 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
5103 X509_LOOKUP_by_alias().
5106 *) Sort openssl functions by name.
5109 *) Get the gendsa program working (hopefully) and add it to app list. Remove
5110 encryption from sample DSA keys (in case anyone is interested the password
5114 *) Make _all_ *_free functions accept a NULL pointer.
5115 [Frans Heymans <fheymans@isaserver.be>]
5117 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
5119 [Anonymous <nobody@replay.com>]
5121 *) s_server should send the CAfile as acceptable CAs, not its own cert.
5122 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
5124 *) Don't blow it for numeric -newkey arguments to apps/req.
5125 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
5127 *) Temp key "for export" tests were wrong in s3_srvr.c.
5128 [Anonymous <nobody@replay.com>]
5130 *) Add prototype for temp key callback functions
5131 SSL_CTX_set_tmp_{rsa,dh}_callback().
5134 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
5135 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
5138 *) X509_name_add_entry() freed the wrong thing after an error.
5139 [Arne Ansper <arne@ats.cyber.ee>]
5141 *) rsa_eay.c would attempt to free a NULL context.
5142 [Arne Ansper <arne@ats.cyber.ee>]
5144 *) BIO_s_socket() had a broken should_retry() on Windoze.
5145 [Arne Ansper <arne@ats.cyber.ee>]
5147 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
5148 [Arne Ansper <arne@ats.cyber.ee>]
5150 *) Make sure the already existing X509_STORE->depth variable is initialized
5151 in X509_STORE_new(), but document the fact that this variable is still
5152 unused in the certificate verification process.
5153 [Ralf S. Engelschall]
5155 *) Fix the various library and apps files to free up pkeys obtained from
5156 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
5159 *) Fix reference counting in X509_PUBKEY_get(). This makes
5160 demos/maurice/example2.c work, amongst others, probably.
5161 [Steve Henson and Ben Laurie]
5163 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
5164 `openssl' and second, the shortcut symlinks for the `openssl <command>'
5165 are no longer created. This way we have a single and consistent command
5166 line interface `openssl <command>', similar to `cvs <command>'.
5167 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
5169 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
5170 BIT STRING wrapper always have zero unused bits.
5173 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
5176 *) Make the top-level INSTALL documentation easier to understand.
5179 *) Makefiles updated to exit if an error occurs in a sub-directory
5180 make (including if user presses ^C) [Paul Sutton]
5182 *) Make Montgomery context stuff explicit in RSA data structure.
5185 *) Fix build order of pem and err to allow for generated pem.h.
5188 *) Fix renumbering bug in X509_NAME_delete_entry().
5191 *) Enhanced the err-ins.pl script so it makes the error library number
5192 global and can add a library name. This is needed for external ASN1 and
5193 other error libraries.
5196 *) Fixed sk_insert which never worked properly.
5199 *) Fix ASN1 macros so they can handle indefinite length construted
5200 EXPLICIT tags. Some non standard certificates use these: they can now
5204 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
5205 into a single doc/ssleay.txt bundle. This way the information is still
5206 preserved but no longer messes up this directory. Now it's new room for
5207 the new set of documenation files.
5208 [Ralf S. Engelschall]
5210 *) SETs were incorrectly DER encoded. This was a major pain, because they
5211 shared code with SEQUENCEs, which aren't coded the same. This means that
5212 almost everything to do with SETs or SEQUENCEs has either changed name or
5213 number of arguments.
5214 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
5216 *) Fix test data to work with the above.
5219 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
5220 was already fixed by Eric for 0.9.1 it seems.
5221 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
5223 *) Autodetect FreeBSD3.
5226 *) Fix various bugs in Configure. This affects the following platforms:
5230 unixware-2.0-pentium
5234 *) Eliminate generated files from CVS. Reorder tests to regenerate files
5235 before they are needed.
5238 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
5242 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
5244 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
5245 changed SSLeay to OpenSSL in version strings.
5246 [Ralf S. Engelschall]
5248 *) Some fixups to the top-level documents.
5251 *) Fixed the nasty bug where rsaref.h was not found under compile-time
5252 because the symlink to include/ was missing.
5253 [Ralf S. Engelschall]
5255 *) Incorporated the popular no-RSA/DSA-only patches
5256 which allow to compile a RSA-free SSLeay.
5257 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
5259 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
5260 when "ssleay" is still not found.
5261 [Ralf S. Engelschall]
5263 *) Added more platforms to Configure: Cray T3E, HPUX 11,
5264 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
5266 *) Updated the README file.
5267 [Ralf S. Engelschall]
5269 *) Added various .cvsignore files in the CVS repository subdirs
5270 to make a "cvs update" really silent.
5271 [Ralf S. Engelschall]
5273 *) Recompiled the error-definition header files and added
5274 missing symbols to the Win32 linker tables.
5275 [Ralf S. Engelschall]
5277 *) Cleaned up the top-level documents;
5278 o new files: CHANGES and LICENSE
5279 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
5280 o merged COPYRIGHT into LICENSE
5281 o removed obsolete TODO file
5282 o renamed MICROSOFT to INSTALL.W32
5283 [Ralf S. Engelschall]
5285 *) Removed dummy files from the 0.9.1b source tree:
5286 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
5287 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
5288 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
5289 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
5290 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
5291 [Ralf S. Engelschall]
5293 *) Added various platform portability fixes.
5296 *) The Genesis of the OpenSSL rpject:
5297 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
5298 Young and Tim J. Hudson created while they were working for C2Net until
5300 [The OpenSSL Project]
5303 Changes between 0.9.0b and 0.9.1b [not released]
5305 *) Updated a few CA certificates under certs/
5308 *) Changed some BIGNUM api stuff.
5311 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
5312 DGUX x86, Linux Alpha, etc.
5315 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
5316 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
5320 *) Add -strparse option to asn1pars program which parses nested
5322 [Dr Stephen Henson <shenson@bigfoot.com>]
5324 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
5327 *) DSA fix for "ca" program.
5330 *) Added "-genkey" option to "dsaparam" program.
5333 *) Added RIPE MD160 (rmd160) message digest.
5336 *) Added -a (all) option to "ssleay version" command.
5339 *) Added PLATFORM define which is the id given to Configure.
5342 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
5345 *) Extended the ASN.1 parser routines.
5348 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
5351 *) Added a BN_CTX to the BN library.
5354 *) Fixed the weak key values in DES library
5357 *) Changed API in EVP library for cipher aliases.
5360 *) Added support for RC2/64bit cipher.
5363 *) Converted the lhash library to the crypto/mem.c functions.
5366 *) Added more recognized ASN.1 object ids.
5369 *) Added more RSA padding checks for SSL/TLS.
5372 *) Added BIO proxy/filter functionality.
5375 *) Added extra_certs to SSL_CTX which can be used
5376 send extra CA certificates to the client in the CA cert chain sending
5377 process. It can be configured with SSL_CTX_add_extra_chain_cert().
5380 *) Now Fortezza is denied in the authentication phase because
5381 this is key exchange mechanism is not supported by SSLeay at all.
5384 *) Additional PKCS1 checks.
5387 *) Support the string "TLSv1" for all TLS v1 ciphers.
5390 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
5391 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
5394 *) Fixed a few memory leaks.
5397 *) Fixed various code and comment typos.
5400 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
5401 bytes sent in the client random.
5402 [Edward Bishop <ebishop@spyglass.com>]