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 +) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
16 so that complete 'Handshake' protocol structures are kept in memory
17 instead of overwriting 'msg_type' and 'length' with 'body' data.
20 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
24 +) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
25 [Massimo Santin via Richard Levitte]
27 +) Major restructuring to the underlying ENGINE code. This includes
28 reduction of linker bloat, separation of pure "ENGINE" manipulation
29 (initialisation, etc) from functionality dealing with implementations
30 of specific crypto iterfaces. This change also introduces integrated
31 support for symmetric ciphers and digest implementations - so ENGINEs
32 can now accelerate these by providing EVP_CIPHER and EVP_MD
33 implementations of their own. This is detailed in crypto/engine/README
34 as it couldn't be adequately described here. However, there are a few
35 API changes worth noting - some RSA, DSA, DH, and RAND functions that
36 were changed in the original introduction of ENGINE code have now
37 reverted back - the hooking from this code to ENGINE is now a good
38 deal more passive and at run-time, operations deal directly with
39 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
40 dereferencing through an ENGINE pointer any more. Also, the ENGINE
41 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
42 they were not being used by the framework as there is no concept of a
43 BIGNUM_METHOD and they could not be generalised to the new
44 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
45 ENGINE_cpy() has been removed as it cannot be consistently defined in
49 +) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
52 +) Change mkdef.pl to sort symbols that get the same entry number,
53 and make sure the automatically generated functions ERR_load_*
54 become part of libeay.num as well.
57 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
58 client receives HelloRequest while in a handshake.
59 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
61 +) New function SSL_renegotiate_pending(). This returns true once
62 renegotiation has been requested (either SSL_renegotiate() call
63 or HelloRequest/ClientHello receveived from the peer) and becomes
64 false once a handshake has been completed.
65 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
66 sends a HelloRequest, but does not ensure that a handshake takes
67 place. SSL_renegotiate_pending() is useful for checking if the
68 client has followed the request.)
71 +) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
72 By default, clients may request session resumption even during
73 renegotiation (if session ID contexts permit); with this option,
74 session resumption is possible only in the first handshake.
77 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
78 should end in 'break', not 'goto end' which circuments various
79 cleanups done in state SSL_ST_OK. But session related stuff
80 must be disabled for SSL_ST_OK in the case that we just sent a
83 Also avoid some overhead by not calling ssl_init_wbio_buffer()
84 before just sending a HelloRequest.
85 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
87 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
88 reveal whether illegal block cipher padding was found or a MAC
89 verification error occured. (Neither SSLerr() codes nor alerts
90 are directly visible to potential attackers, but the information
91 may leak via logfiles.)
93 Similar changes are not required for the SSL 2.0 implementation
94 because the number of padding bytes is sent in clear for SSL 2.0,
95 and the extra bytes are just ignored. However ssl/s2_pkt.c
96 failed to verify that the purported number of padding bytes is in
100 +) Add some demos for certificate and certificate request creation.
103 +) Make maximum certificate chain size accepted from the peer application
104 settable (SSL*_get/set_max_cert_list()), as proposed by
105 "Douglas E. Engert" <deengert@anl.gov>.
108 +) Add support for shared libraries for Unixware-7 and support including
109 shared libraries for OpenUNIX-8 (Boyd Lynn Gerber <gerberb@zenez.com>).
112 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
113 'wristwatch attack' using huge encoding parameters (cf.
114 James H. Manger's CRYPTO 2001 paper). Note that the
115 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
116 encoding parameters and hence was not vulnerable.
119 +) Add a "destroy" handler to ENGINEs that allows structural cleanup to
120 be done prior to destruction. Use this to unload error strings from
121 ENGINEs that load their own error strings. NB: This adds two new API
122 functions to "get" and "set" this destroy handler in an ENGINE.
125 +) Alter all existing ENGINE implementations (except "openssl" and
126 "openbsd") to dynamically instantiate their own error strings. This
127 makes them more flexible to be built both as statically-linked ENGINEs
128 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
129 Also, add stub code to each that makes building them as self-contained
130 shared-libraries easier (see README.ENGINE).
133 +) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
134 implementations into applications that are completely implemented in
135 self-contained shared-libraries. The "dynamic" ENGINE exposes control
136 commands that can be used to configure what shared-library to load and
137 to control aspects of the way it is handled. Also, made an update to
138 the README.ENGINE file that brings its information up-to-date and
139 provides some information and instructions on the "dynamic" ENGINE
140 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
144 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
146 *) Make it possible to unload ranges of ERR strings with a new
147 "ERR_unload_strings" function.
150 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
151 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
152 followed by modular reduction.
153 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
155 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
156 equivalent based on BN_pseudo_rand() instead of BN_rand().
159 +) Add a copy() function to EVP_MD.
162 +) Make EVP_MD routines take a context pointer instead of just the
163 md_data void pointer.
166 +) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
167 that the digest can only process a single chunk of data
168 (typically because it is provided by a piece of
169 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
170 is only going to provide a single chunk of data, and hence the
171 framework needn't accumulate the data for oneshot drivers.
174 +) As with "ERR", make it possible to replace the underlying "ex_data"
175 functions. This change also alters the storage and management of global
176 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
177 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
178 index counters. The API functions that use this state have been changed
179 to take a "class_index" rather than pointers to the class's local STACK
180 and counter, and there is now an API function to dynamically create new
181 classes. This centralisation allows us to (a) plug a lot of the
182 thread-safety problems that existed, and (b) makes it possible to clean
183 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
184 such data would previously have always leaked in application code and
185 workarounds were in place to make the memory debugging turn a blind eye
186 to it. Application code that doesn't use this new function will still
187 leak as before, but their memory debugging output will announce it now
188 rather than letting it slide.
190 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
191 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
192 has a return value to indicate success or failure.
195 +) Make it possible to replace the underlying "ERR" functions such that the
196 global state (2 LHASH tables and 2 locks) is only used by the "default"
197 implementation. This change also adds two functions to "get" and "set"
198 the implementation prior to it being automatically set the first time
199 any other ERR function takes place. Ie. an application can call "get",
200 pass the return value to a module it has just loaded, and that module
201 can call its own "set" function using that value. This means the
202 module's "ERR" operations will use (and modify) the error state in the
203 application and not in its own statically linked copy of OpenSSL code.
206 +) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
207 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
208 the operation, and provides a more encapsulated way for external code
209 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
210 to use these functions rather than manually incrementing the counts.
212 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
215 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
216 This function was broken, as the check for a new client hello message
217 to handle SGC did not allow these large messages.
218 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
221 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
224 +) Add EVP test program.
227 +) Add symmetric cipher support to ENGINE. Expect the API to change!
230 +) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
231 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
232 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
233 These allow a CRL to be built without having to access X509_CRL fields
234 directly. Modify 'ca' application to use new functions.
237 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
238 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
241 *) Rework the configuration and shared library support for Tru64 Unix.
242 The configuration part makes use of modern compiler features and
243 still retains old compiler behavior for those that run older versions
244 of the OS. The shared library support part includes a variant that
245 uses the RPATH feature, and is available through the special
246 configuration target "alpha-cc-rpath", which will never be selected
248 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
250 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
251 with the same message size as in ssl3_get_certificate_request().
252 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
253 messages might inadvertently be reject as too long.
254 [Petr Lampa <lampa@fee.vutbr.cz>]
256 +) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
257 bug workarounds. Rollback attack detection is a security feature.
258 The problem will only arise on OpenSSL servers when TLSv1 is not
259 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
260 Software authors not wanting to support TLSv1 will have special reasons
261 for their choice and can explicitly enable this option.
262 [Bodo Moeller, Lutz Jaenicke]
264 +) Rationalise EVP so it can be extended: don't include a union of
265 cipher/digest structures, add init/cleanup functions. This also reduces
266 the number of header dependencies.
269 +) Make DES key schedule conform to the usual scheme, as well as
270 correcting its structure. This means that calls to DES functions
271 now have to pass a pointer to a des_key_schedule instead of a
272 plain des_key_schedule (which was actually always a pointer
276 +) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
279 *) Modified SSL library such that the verify_callback that has been set
280 specificly for an SSL object with SSL_set_verify() is actually being
281 used. Before the change, a verify_callback set with this function was
282 ignored and the verify_callback() set in the SSL_CTX at the time of
283 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
284 to allow the necessary settings.
287 +) Initial reduction of linker bloat: the use of some functions, such as
288 PEM causes large amounts of unused functions to be linked in due to
289 poor organisation. For example pem_all.c contains every PEM function
290 which has a knock on effect of linking in large amounts of (unused)
291 ASN1 code. Grouping together similar functions and splitting unrelated
292 functions prevents this.
295 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
296 explicitly to NULL, as at least on Solaris 8 this seems not always to be
297 done automatically (in contradiction to the requirements of the C
298 standard). This made problems when used from OpenSSH.
301 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
302 dh->length and always used
304 BN_rand_range(priv_key, dh->p).
306 BN_rand_range() is not necessary for Diffie-Hellman, and this
307 specific range makes Diffie-Hellman unnecessarily inefficient if
308 dh->length (recommended exponent length) is much smaller than the
309 length of dh->p. We could use BN_rand_range() if the order of
310 the subgroup was stored in the DH structure, but we only have
315 BN_rand(priv_key, l, ...)
317 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
323 RSA_eay_public_encrypt
324 RSA_eay_private_decrypt
325 RSA_eay_private_encrypt (signing)
326 RSA_eay_public_decrypt (signature verification)
328 (default implementations for RSA_public_encrypt,
329 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
330 always reject numbers >= n.
333 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
334 to synchronize access to 'locking_thread'. This is necessary on
335 systems where access to 'locking_thread' (an 'unsigned long'
336 variable) is not atomic.
339 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
340 *before* setting the 'crypto_lock_rand' flag. The previous code had
341 a race condition if 0 is a valid thread ID.
342 [Travis Vitek <vitek@roguewave.com>]
344 +) Cleanup of EVP macros.
347 +) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
351 +) Add initial OCSP responder support to ocsp application. The
352 revocation information is handled using the text based index
353 use by the ca application. The responder can either handle
354 requests generated internally, supplied in files (for example
355 via a CGI script) or using an internal minimal server.
358 +) Add configuration choices to get zlib compression for TLS.
361 +) Changes to Kerberos SSL for RFC 2712 compliance:
362 1. Implemented real KerberosWrapper, instead of just using
363 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
364 2. Implemented optional authenticator field of KerberosWrapper.
366 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
367 and authenticator structs; see crypto/krb5/.
369 Generalized Kerberos calls to support multiple Kerberos libraries.
370 [Vern Staats <staatsvr@asc.hpc.mil>,
371 Jeffrey Altman <jaltman@columbia.edu>
374 +) Cause 'openssl speed' to use fully hard-coded DSA keys as it
375 already does with RSA. testdsa.h now has 'priv_key/pub_key'
376 values for each of the key sizes rather than having just
377 parameters (and 'speed' generating keys each time).
380 -) OpenSSL 0.9.6b released [9 July 2001]
382 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
383 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
384 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
385 PRNG state recovery was possible based on the output of
386 one PRNG request appropriately sized to gain knowledge on
387 'md' followed by enough consecutive 1-byte PRNG requests
388 to traverse all of 'state'.
390 1. When updating 'md_local' (the current thread's copy of 'md')
391 during PRNG output generation, hash all of the previous
392 'md_local' value, not just the half used for PRNG output.
394 2. Make the number of bytes from 'state' included into the hash
395 independent from the number of PRNG bytes requested.
397 The first measure alone would be sufficient to avoid
398 Markku-Juhani's attack. (Actually it had never occurred
399 to me that the half of 'md_local' used for chaining was the
400 half from which PRNG output bytes were taken -- I had always
401 assumed that the secret half would be used.) The second
402 measure makes sure that additional data from 'state' is never
403 mixed into 'md_local' in small portions; this heuristically
404 further strengthens the PRNG.
407 +) Speed up EVP routines.
410 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
411 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
412 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
413 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
415 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
416 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
417 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
420 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
422 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
425 *) Fix crypto/bn/asm/mips3.s.
428 *) When only the key is given to "enc", the IV is undefined. Print out
429 an error message in this case.
432 +) Added the OS2-EMX target.
433 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
435 +) Rewrite apps to use NCONF routines instead of the old CONF. New functions
436 to support NCONF routines in extension code. New function CONF_set_nconf()
437 to allow functions which take an NCONF to also handle the old LHASH
438 structure: this means that the old CONF compatible routines can be
439 retained (in particular wrt extensions) without having to duplicate the
440 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
443 *) Handle special case when X509_NAME is empty in X509 printing routines.
446 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
447 positive and less than q.
450 +) Enhance the general user interface with mechanisms for inner control
451 and with possibilities to have yes/no kind of prompts.
454 +) Change all calls to low level digest routines in the library and
455 applications to use EVP. Add missing calls to HMAC_cleanup() and
456 don't assume HMAC_CTX can be copied using memcpy().
457 [Verdon Walker <VWalker@novell.com>, Steve Henson]
459 +) Add the possibility to control engines through control names but with
460 arbitrary arguments instead of just a string.
461 Change the key loaders to take a UI_METHOD instead of a callback
462 function pointer. NOTE: this breaks binary compatibility with earlier
463 versions of OpenSSL [engine].
464 Adapt the nCipher code for these new conditions and add a card insertion
468 +) Enhance the general user interface with mechanisms to better support
469 dialog box interfaces, application-defined prompts, the possibility
470 to use defaults (for example default passwords from somewhere else)
471 and interrupts/cancellations.
474 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
475 used: it isn't thread safe and the add_lock_callback should handle
477 [Paul Rose <Paul.Rose@bridge.com>]
479 *) Verify that incoming data obeys the block size in
480 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
483 +) Tidy up PKCS#12 attribute handling. Add support for the CSP name
484 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
488 [Ulf Möller, Bodo Möller]
490 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
491 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
492 when fixing the server behaviour for backwards-compatible 'client
493 hello' messages. (Note that the attack is impractical against
494 SSL 3.0 and TLS 1.0 anyway because length and version checking
495 means that the probability of guessing a valid ciphertext is
496 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
499 Before 0.9.5, the countermeasure (hide the error by generating a
500 random 'decryption result') did not work properly because
501 ERR_clear_error() was missing, meaning that SSL_get_error() would
502 detect the supposedly ignored error.
504 Both problems are now fixed.
507 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
508 (previously it was 1024).
511 +) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
512 tidy up some unnecessarily weird code in 'sk_new()').
513 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
515 +) Change the key loading routines for ENGINEs to use the same kind
516 callback (pem_password_cb) as all other routines that need this
520 *) Fix for compatibility mode trust settings: ignore trust settings
521 unless some valid trust or reject settings are present.
524 *) Fix for blowfish EVP: its a variable length cipher.
527 +) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
528 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
529 than this minimum value is recommended.
532 +) New random seeder for OpenVMS, using the system process statistics
533 that are easily reachable.
536 +) Windows apparently can't transparently handle global
537 variables defined in DLLs. Initialisations such as:
539 const ASN1_ITEM *it = &ASN1_INTEGER_it;
541 wont compile. This is used by the any applications that need to
542 declare their own ASN1 modules. This was fixed by adding the option
543 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
544 needed for static libraries under Win32.
547 +) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
548 setting of purpose and trust fields. New X509_STORE trust and
549 purpose functions and tidy up setting in other SSL functions.
552 +) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
553 structure. These are inherited by X509_STORE_CTX when it is
554 initialised. This allows various defaults to be set in the
555 X509_STORE structure (such as flags for CRL checking and custom
556 purpose or trust settings) for functions which only use X509_STORE_CTX
557 internally such as S/MIME.
559 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
560 trust settings if they are not set in X509_STORE. This allows X509_STORE
561 purposes and trust (in S/MIME for example) to override any set by default.
563 Add command line options for CRL checking to smime, s_client and s_server
567 +) Initial CRL based revocation checking. If the CRL checking flag(s)
568 are set then the CRL is looked up in the X509_STORE structure and
569 its validity and signature checked, then if the certificate is found
570 in the CRL the verify fails with a revoked error.
572 Various new CRL related callbacks added to X509_STORE_CTX structure.
574 Command line options added to 'verify' application to support this.
576 This needs some additional work, such as being able to handle multiple
577 CRLs with different times, extension based lookup (rather than just
578 by subject name) and ultimately more complete V2 CRL extension
582 +) Add a general user interface API (crypto/ui/). This is designed
583 to replace things like des_read_password and friends (backward
584 compatibility functions using this new API are provided).
585 The purpose is to remove prompting functions from the DES code
586 section as well as provide for prompting through dialog boxes in
587 a window system and the like.
590 *) Fix various bugs related to DSA S/MIME verification. Handle missing
591 parameters in DSA public key structures and return an error in the
592 DSA routines if parameters are absent.
595 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
596 in the current directory if neither $RANDFILE nor $HOME was set.
597 RAND_file_name() in 0.9.6a returned NULL in this case. This has
598 caused some confusion to Windows users who haven't defined $HOME.
599 Thus RAND_file_name() is changed again: e_os.h can define a
600 DEFAULT_HOME, which will be used if $HOME is not set.
601 For Windows, we use "C:"; on other platforms, we still require
602 environment variables.
604 +) Add "ex_data" support to ENGINE so implementations can add state at a
605 per-structure level rather than having to store it globally.
608 +) Make it possible for ENGINE structures to be copied when retrieved by
609 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
610 This causes the "original" ENGINE structure to act like a template,
611 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
612 operational state can be localised to each ENGINE structure, despite the
613 fact they all share the same "methods". New ENGINE structures returned in
614 this case have no functional references and the return value is the single
615 structural reference. This matches the single structural reference returned
616 by ENGINE_by_id() normally, when it is incremented on the pre-existing
620 +) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
621 needs to match any other type at all we need to manually clear the
625 +) Changes to the "openssl engine" utility to include;
626 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
627 about an ENGINE's available control commands.
628 - executing control commands from command line arguments using the
629 '-pre' and '-post' switches. '-post' is only used if '-t' is
630 specified and the ENGINE is successfully initialised. The syntax for
631 the individual commands are colon-separated, for example;
632 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
635 +) New dynamic control command support for ENGINEs. ENGINEs can now
636 declare their own commands (numbers), names (strings), descriptions,
637 and input types for run-time discovery by calling applications. A
638 subset of these commands are implicitly classed as "executable"
639 depending on their input type, and only these can be invoked through
640 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
641 can be based on user input, config files, etc). The distinction is
642 that "executable" commands cannot return anything other than a boolean
643 result and can only support numeric or string input, whereas some
644 discoverable commands may only be for direct use through
645 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
646 pointers, or other custom uses. The "executable" commands are to
647 support parameterisations of ENGINE behaviour that can be
648 unambiguously defined by ENGINEs and used consistently across any
649 OpenSSL-based application. Commands have been added to all the
650 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
651 control over shared-library paths without source code alterations.
654 +) Changed all ENGINE implementations to dynamically allocate their
655 ENGINEs rather than declaring them statically. Apart from this being
656 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
657 this also allows the implementations to compile without using the
658 internal engine_int.h header.
661 +) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
662 'const' value. Any code that should be able to modify a RAND_METHOD
663 should already have non-const pointers to it (ie. they should only
664 modify their own ones).
667 +) Made a variety of little tweaks to the ENGINE code.
668 - "atalla" and "ubsec" string definitions were moved from header files
669 to C code. "nuron" string definitions were placed in variables
670 rather than hard-coded - allowing parameterisation of these values
671 later on via ctrl() commands.
672 - Removed unused "#if 0"'d code.
673 - Fixed engine list iteration code so it uses ENGINE_free() to release
674 structural references.
675 - Constified the RAND_METHOD element of ENGINE structures.
676 - Constified various get/set functions as appropriate and added
677 missing functions (including a catch-all ENGINE_cpy that duplicates
678 all ENGINE values onto a new ENGINE except reference counts/state).
679 - Removed NULL parameter checks in get/set functions. Setting a method
680 or function to NULL is a way of cancelling out a previously set
681 value. Passing a NULL ENGINE parameter is just plain stupid anyway
682 and doesn't justify the extra error symbols and code.
683 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
684 flags from engine_int.h to engine.h.
685 - Changed prototypes for ENGINE handler functions (init(), finish(),
686 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
689 *) Move 'if (!initialized) RAND_poll()' into regions protected by
690 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
691 having multiple threads call RAND_poll() concurrently.
694 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
695 combination of a flag and a thread ID variable.
696 Otherwise while one thread is in ssleay_rand_bytes (which sets the
697 flag), *other* threads can enter ssleay_add_bytes without obeying
698 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
699 that they do not hold after the first thread unsets add_do_not_lock).
702 +) Implement binary inversion algorithm for BN_mod_inverse in addition
703 to the algorithm using long division. The binary algorithm can be
704 used only if the modulus is odd. On 32-bit systems, it is faster
705 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
706 roughly 5-15% for 256-bit moduli), so we use it only for moduli
707 up to 450 bits. In 64-bit environments, the binary algorithm
708 appears to be advantageous for much longer moduli; here we use it
709 for moduli up to 2048 bits.
712 *) Change bctest again: '-x' expressions are not available in all
716 -) OpenSSL 0.9.6a released [5 Apr 2001]
718 *) Fix a couple of memory leaks in PKCS7_dataDecode()
719 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
721 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
722 the default extension for executables, if any. Also, make the perl
723 scripts that use symlink() to test if it really exists and use "cp"
724 if it doesn't. All this made OpenSSL compilable and installable in
728 +) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
729 could not support the combine flag in choice fields.
732 -) Fix for asn1_GetSequence() for indefinite length constructed data.
733 If SEQUENCE is length is indefinite just set c->slen to the total
734 amount of data available.
735 [Steve Henson, reported by shige@FreeBSD.org]
736 [This change does not apply to 0.9.7.]
738 *) Change bctest to avoid here-documents inside command substitution
739 (workaround for FreeBSD /bin/sh bug).
740 For compatibility with Ultrix, avoid shell functions (introduced
741 in the bctest version that searches along $PATH).
744 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
745 with des_encrypt() defined on some operating systems, like Solaris
749 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
750 On the Importance of Eliminating Errors in Cryptographic
751 Computations, J. Cryptology 14 (2001) 2, 101-119,
752 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
755 *) MIPS assembler BIGNUM division bug fix.
758 *) Disabled incorrect Alpha assembler code.
761 -) Fix PKCS#7 decode routines so they correctly update the length
762 after reading an EOC for the EXPLICIT tag.
764 [This change does not apply to 0.9.7.]
766 *) Fix bug in PKCS#12 key generation routines. This was triggered
767 if a 3DES key was generated with a 0 initial byte. Include
768 PKCS12_BROKEN_KEYGEN compilation option to retain the old
769 (but broken) behaviour.
772 *) Enhance bctest to search for a working bc along $PATH and print
774 [Tim Rice <tim@multitalents.net> via Richard Levitte]
776 +) Add a 'copy_extensions' option to the 'ca' utility. This copies
777 extensions from a certificate request to the certificate.
780 +) Allow multiple 'certopt' and 'nameopt' options to be separated
781 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
782 file: this allows the display of the certificate about to be
783 signed to be customised, to allow certain fields to be included
784 or excluded and extension details. The old system didn't display
785 multicharacter strings properly, omitted fields not in the policy
786 and couldn't display additional details such as extensions.
789 *) Fix memory leaks in err.c: free err_data string if necessary;
790 don't write to the wrong index in ERR_set_error_data.
793 +) Function EC_POINTs_mul for simultaneous scalar multiplication
794 of an arbitrary number of elliptic curve points, optionally
795 including the generator defined for the EC_GROUP.
796 EC_POINT_mul is a simple wrapper function for the typical case
797 that the point list has just one item (besides the optional
801 +) First EC_METHODs for curves over GF(p):
803 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
804 operations and provides various method functions that can also
805 operate with faster implementations of modular arithmetic.
807 EC_GFp_mont_method() reuses most functions that are part of
808 EC_GFp_simple_method, but uses Montgomery arithmetic.
810 [Bodo Moeller; point addition and point doubling
811 implementation directly derived from source code provided by
812 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
814 +) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
817 Curves are EC_GROUP objects (with an optional group generator)
818 based on EC_METHODs that are built into the library.
820 Points are EC_POINT objects based on EC_GROUP objects.
822 Most of the framework would be able to handle curves over arbitrary
823 finite fields, but as there are no obvious types for fields other
824 than GF(p), some functions are limited to that for now.
827 +) Add the -HTTP option to s_server. It is similar to -WWW, but requires
828 that the file contains a complete HTTP response.
831 +) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
832 change the def and num file printf format specifier from "%-40sXXX"
833 to "%-39s XXX". The latter will always guarantee a space after the
834 field while the former will cause them to run together if the field
835 is 40 of more characters long.
838 +) Constify the cipher and digest 'method' functions and structures
839 and modify related functions to take constant EVP_MD and EVP_CIPHER
843 *) Implement ssl23_peek (analogous to ssl23_read), which previously
847 *) Replace rdtsc with _emit statements for VC++ version 5.
848 [Jeremy Cooper <jeremy@baymoo.org>]
850 +) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
851 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
854 +) Modify EVP_Digest*() routines so they now return values. Although the
855 internal software routines can never fail additional hardware versions
859 +) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
861 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
862 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
865 ERR_R_NESTED_ASN1_ERROR
867 ERR_R_MISSING_ASN1_EOS
868 were 4 .. 9, conflicting with
869 ERR_LIB_RSA (= ERR_R_RSA_LIB)
871 ERR_LIB_PEM (= ERR_R_PEM_LIB).
872 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
874 Add new error code 'ERR_R_INTERNAL_ERROR'.
877 +) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
881 +) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
882 sets the subject name for a new request or supersedes the
883 subject name in a given request. Formats that can be parsed are
884 'CN=Some Name, OU=myOU, C=IT'
886 'CN=Some Name/OU=myOU/C=IT'.
888 Add options '-batch' and '-verbose' to 'openssl req'.
889 [Massimiliano Pala <madwolf@hackmasters.net>]
891 *) Make it possible to reuse SSLv2 sessions.
894 +) Introduce the possibility to access global variables through
895 functions on platform were that's the best way to handle exporting
896 global variables in shared libraries. To enable this functionality,
897 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
898 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
899 is normally done by Configure or something similar).
901 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
902 in the source file (foo.c) like this:
904 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
905 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
907 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
908 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
910 OPENSSL_DECLARE_GLOBAL(int,foo);
911 #define foo OPENSSL_GLOBAL_REF(foo)
912 OPENSSL_DECLARE_GLOBAL(double,bar);
913 #define bar OPENSSL_GLOBAL_REF(bar)
915 The #defines are very important, and therefore so is including the
916 header file everywhere where the defined globals are used.
918 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
919 of ASN.1 items, but that structure is a bit different.
921 The largest change is in util/mkdef.pl which has been enhanced with
922 better and easier to understand logic to choose which symbols should
923 go into the Windows .def files as well as a number of fixes and code
924 cleanup (among others, algorithm keywords are now sorted
925 lexicographically to avoid constant rewrites).
928 *) In copy_email() check for >= 0 as a return value for
929 X509_NAME_get_index_by_NID() since 0 is a valid index.
930 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
932 +) In BN_div() keep a copy of the sign of 'num' before writing the
933 result to 'rm' because if rm==num the value will be overwritten
934 and produce the wrong result if 'num' is negative: this caused
935 problems with BN_mod() and BN_nnmod().
938 +) Function OCSP_request_verify(). This checks the signature on an
939 OCSP request and verifies the signer certificate. The signer
940 certificate is just checked for a generic purpose and OCSP request
944 +) Add OCSP_check_validity() function to check the validity of OCSP
945 responses. OCSP responses are prepared in real time and may only
946 be a few seconds old. Simply checking that the current time lies
947 between thisUpdate and nextUpdate max reject otherwise valid responses
948 caused by either OCSP responder or client clock inaccuracy. Instead
949 we allow thisUpdate and nextUpdate to fall within a certain period of
950 the current time. The age of the response can also optionally be
951 checked. Two new options -validity_period and -status_age added to
955 +) If signature or public key algorithm is unrecognized print out its
956 OID rather that just UNKNOWN.
959 *) Avoid coredump with unsupported or invalid public keys by checking if
960 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
961 PKCS7_verify() fails with non detached data.
964 +) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
965 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
966 ID to be generated from the issuer certificate alone which can then be
967 passed to OCSP_id_issuer_cmp().
970 +) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
971 ASN1 modules to export functions returning ASN1_ITEM pointers
972 instead of the ASN1_ITEM structures themselves. This adds several
973 new macros which allow the underlying ASN1 function/structure to
974 be accessed transparently. As a result code should not use ASN1_ITEM
975 references directly (such as &X509_it) but instead use the relevant
976 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
977 use of the new ASN1 code on platforms where exporting structures
978 is problematical (for example in shared libraries) but exporting
979 functions returning pointers to structures is not.
982 +) Add support for overriding the generation of SSL/TLS session IDs.
983 These callbacks can be registered either in an SSL_CTX or per SSL.
984 The purpose of this is to allow applications to control, if they wish,
985 the arbitrary values chosen for use as session IDs, particularly as it
986 can be useful for session caching in multiple-server environments. A
987 command-line switch for testing this (and any client code that wishes
988 to use such a feature) has been added to "s_server".
989 [Geoff Thorpe, Lutz Jaenicke]
991 +) Modify mkdef.pl to recognise and parse preprocessor conditionals
992 of the form '#if defined(...) || defined(...) || ...' and
993 '#if !defined(...) && !defined(...) && ...'. This also avoids
994 the growing number of special cases it was previously handling.
997 *) Don't use getenv in library functions when run as setuid/setgid.
998 New function OPENSSL_issetugid().
1001 +) Make all configuration macros available for application by making
1002 sure they are available in opensslconf.h, by giving them names starting
1003 with "OPENSSL_" to avoid conflicts with other packages and by making
1004 sure e_os2.h will cover all platform-specific cases together with
1006 Additionally, it is now possible to define configuration/platform-
1007 specific names (called "system identities"). In the C code, these
1008 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
1009 macro with the name beginning with "OPENSSL_SYS_", which is determined
1010 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
1014 +) New option -set_serial to 'req' and 'x509' this allows the serial
1015 number to use to be specified on the command line. Previously self
1016 signed certificates were hard coded with serial number 0 and the
1017 CA options of 'x509' had to use a serial number in a file which was
1021 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
1022 due to incorrect handling of multi-threading:
1024 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
1026 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
1028 3. Count how many times MemCheck_off() has been called so that
1029 nested use can be treated correctly. This also avoids
1030 inband-signalling in the previous code (which relied on the
1031 assumption that thread ID 0 is impossible).
1034 +) New options to 'ca' utility to support V2 CRL entry extensions.
1035 Currently CRL reason, invalidity date and hold instruction are
1036 supported. Add new CRL extensions to V3 code and some new objects.
1039 *) Add "-rand" option also to s_client and s_server.
1042 +) New function EVP_CIPHER_CTX_set_padding() this is used to
1043 disable standard block padding (aka PKCS#5 padding) in the EVP
1044 API, which was previously mandatory. This means that the data is
1045 not padded in any way and so the total length much be a multiple
1046 of the block size, otherwise an error occurs.
1049 +) Initial (incomplete) OCSP SSL support.
1052 *) Fix CPU detection on Irix 6.x.
1053 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
1054 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
1056 +) New function OCSP_parse_url(). This splits up a URL into its host,
1057 port and path components: primarily to parse OCSP URLs. New -url
1058 option to ocsp utility.
1061 +) New nonce behavior. The return value of OCSP_check_nonce() now
1062 reflects the various checks performed. Applications can decide
1063 whether to tolerate certain situations such as an absent nonce
1064 in a response when one was present in a request: the ocsp application
1065 just prints out a warning. New function OCSP_add1_basic_nonce()
1066 this is to allow responders to include a nonce in a response even if
1067 the request is nonce-less.
1070 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
1073 [This change does not apply to 0.9.7.]
1075 *) Use the cached encoding of an X509_NAME structure rather than
1076 copying it. This is apparently the reason for the libsafe "errors"
1077 but the code is actually correct.
1080 +) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
1081 skipped when using openssl x509 multiple times on a single input file,
1082 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
1085 +) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
1086 set string type: to handle setting ASN1_TIME structures. Fix ca
1087 utility to correctly initialize revocation date of CRLs.
1090 +) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
1091 the clients preferred ciphersuites and rather use its own preferences.
1092 Should help to work around M$ SGC (Server Gated Cryptography) bug in
1093 Internet Explorer by ensuring unchanged hash method during stepup.
1094 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
1097 +) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
1098 to aes and add a new 'exist' option to print out symbols that don't
1102 +) Additional options to ocsp utility to allow flags to be set and
1103 additional certificates supplied.
1106 +) Add the option -VAfile to 'openssl ocsp', so the user can give the
1107 OCSP client a number of certificate to only verify the response
1111 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
1112 Bleichenbacher's DSA attack.
1113 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
1114 to be set and top=0 forces the highest bit to be set; top=-1 is new
1115 and leaves the highest bit random.
1116 [Ulf Moeller, Bodo Moeller]
1118 +) Update Rijndael code to version 3.0 and change EVP AES ciphers to
1119 handle the new API. Currently only ECB, CBC modes supported. Add new
1120 AES OIDs. Add TLS AES ciphersuites as described in the "AES Ciphersuites
1121 for TLS" draft-ietf-tls-ciphersuite-03.txt.
1122 [Ben Laurie, Steve Henson]
1124 *) In the NCONF_...-based implementations for CONF_... queries
1125 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
1126 a temporary CONF structure with the data component set to NULL
1127 (which gives segmentation faults in lh_retrieve).
1128 Instead, use NULL for the CONF pointer in CONF_get_string and
1129 CONF_get_number (which may use environment variables) and directly
1130 return NULL from CONF_get_section.
1133 *) Fix potential buffer overrun for EBCDIC.
1136 +) New function OCSP_copy_nonce() to copy nonce value (if present) from
1137 request to response.
1140 +) Functions for OCSP responders. OCSP_request_onereq_count(),
1141 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
1142 extract information from a certificate request. OCSP_response_create()
1143 creates a response and optionally adds a basic response structure.
1144 OCSP_basic_add1_status() adds a complete single response to a basic
1145 response and returns the OCSP_SINGLERESP structure just added (to allow
1146 extensions to be included for example). OCSP_basic_add1_cert() adds a
1147 certificate to a basic response and OCSP_basic_sign() signs a basic
1148 response with various flags. New helper functions ASN1_TIME_check()
1149 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
1150 (converts ASN1_TIME to GeneralizedTime).
1153 +) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
1154 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
1155 structure from a certificate. X509_pubkey_digest() digests the public_key
1156 contents: this is used in various key identifiers.
1159 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
1160 keyUsage if basicConstraints absent for a CA.
1163 *) Make SMIME_write_PKCS7() write mail header values with a format that
1164 is more generally accepted (no spaces before the semicolon), since
1165 some programs can't parse those values properly otherwise. Also make
1166 sure BIO's that break lines after each write do not create invalid
1170 +) Make sk_sort() tolerate a NULL argument.
1171 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
1173 +) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
1174 passed by the function are trusted implicitly. If any of them signed the
1175 response then it is assumed to be valid and is not verified.
1178 -) Make the CRL encoding routines work with empty SEQUENCE OF. The
1179 macros previously used would not encode an empty SEQUENCE OF
1180 and break the signature.
1182 [This change does not apply to 0.9.7.]
1184 *) Zero the premaster secret after deriving the master secret in
1188 +) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
1189 to data. This was previously part of the PKCS7 ASN1 code. This
1190 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
1191 [Steve Henson, reported by Kenneth R. Robinette
1192 <support@securenetterm.com>]
1194 +) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
1195 routines: without these tracing memory leaks is very painful.
1196 Fix leaks in PKCS12 and PKCS7 routines.
1199 *) Add some EVP_add_digest_alias registrations (as found in
1200 OpenSSL_add_all_digests()) to SSL_library_init()
1201 aka OpenSSL_add_ssl_algorithms(). This provides improved
1202 compatibility with peers using X.509 certificates
1203 with unconventional AlgorithmIdentifier OIDs.
1206 *) Fix for Irix with NO_ASM.
1207 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
1209 *) ./config script fixes.
1210 [Ulf Moeller, Richard Levitte]
1212 +) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
1213 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
1214 effectively meant GeneralizedTime would never be used. Now it
1215 is initialised to -1 but X509_time_adj() now has to check the value
1216 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
1217 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
1218 [Steve Henson, reported by Kenneth R. Robinette
1219 <support@securenetterm.com>]
1221 +) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
1222 result in a zero length in the ASN1_INTEGER structure which was
1223 not consistent with the structure when d2i_ASN1_INTEGER() was used
1224 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
1225 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
1226 where it did not print out a minus for negative ASN1_INTEGER.
1229 *) Fix 'openssl passwd -1'.
1232 +) Add summary printout to ocsp utility. The various functions which
1233 convert status values to strings have been renamed to:
1234 OCSP_response_status_str(), OCSP_cert_status_str() and
1235 OCSP_crl_reason_str() and are no longer static. New options
1236 to verify nonce values and to disable verification. OCSP response
1237 printout format cleaned up.
1240 +) Add additional OCSP certificate checks. These are those specified
1241 in RFC2560. This consists of two separate checks: the CA of the
1242 certificate being checked must either be the OCSP signer certificate
1243 or the issuer of the OCSP signer certificate. In the latter case the
1244 OCSP signer certificate must contain the OCSP signing extended key
1245 usage. This check is performed by attempting to match the OCSP
1246 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
1247 in the OCSP_CERTID structures of the response.
1250 +) Initial OCSP certificate verification added to OCSP_basic_verify()
1251 and related routines. This uses the standard OpenSSL certificate
1252 verify routines to perform initial checks (just CA validity) and
1253 to obtain the certificate chain. Then additional checks will be
1254 performed on the chain. Currently the root CA is checked to see
1255 if it is explicitly trusted for OCSP signing. This is used to set
1256 a root CA as a global signing root: that is any certificate that
1257 chains to that CA is an acceptable OCSP signing certificate.
1260 +) New '-extfile ...' option to 'openssl ca' for reading X.509v3
1261 extensions from a separate configuration file.
1262 As when reading extensions from the main configuration file,
1263 the '-extensions ...' option may be used for specifying the
1265 [Massimiliano Pala <madwolf@comune.modena.it>]
1267 *) Change PKCS12_key_gen_asc() so it can cope with non null
1268 terminated strings whose length is passed in the passlen
1269 parameter, for example from PEM callbacks. This was done
1270 by adding an extra length parameter to asc2uni().
1271 [Steve Henson, reported by <oddissey@samsung.co.kr>]
1273 +) New OCSP utility. Allows OCSP requests to be generated or
1274 read. The request can be sent to a responder and the output
1275 parsed, outputed or printed in text form. Not complete yet:
1276 still needs to check the OCSP response validity.
1279 +) New subcommands for 'openssl ca':
1280 'openssl ca -status <serial>' prints the status of the cert with
1281 the given serial number (according to the index file).
1282 'openssl ca -updatedb' updates the expiry status of certificates
1284 [Massimiliano Pala <madwolf@comune.modena.it>]
1286 +) New '-newreq-nodes' command option to CA.pl. This is like
1287 '-newreq', but calls 'openssl req' with the '-nodes' option
1288 so that the resulting key is not encrypted.
1289 [Damien Miller <djm@mindrot.org>]
1291 +) New configuration for the GNU Hurd.
1292 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
1294 +) Initial code to implement OCSP basic response verify. This
1295 is currently incomplete. Currently just finds the signer's
1296 certificate and verifies the signature on the response.
1299 +) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
1300 value of OPENSSLDIR. This is available via the new '-d' option
1301 to 'openssl version', and is also included in 'openssl version -a'.
1304 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
1305 call failed, free the DSA structure.
1308 +) Allowing defining memory allocation callbacks that will be given
1309 file name and line number information in additional arguments
1310 (a const char* and an int). The basic functionality remains, as
1311 well as the original possibility to just replace malloc(),
1312 realloc() and free() by functions that do not know about these
1313 additional arguments. To register and find out the current
1314 settings for extended allocation functions, the following
1315 functions are provided:
1317 CRYPTO_set_mem_ex_functions
1318 CRYPTO_set_locked_mem_ex_functions
1319 CRYPTO_get_mem_ex_functions
1320 CRYPTO_get_locked_mem_ex_functions
1322 These work the same way as CRYPTO_set_mem_functions and friends.
1323 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
1324 extended allocation function is enabled.
1325 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
1326 a conventional allocation function is enabled.
1327 [Richard Levitte, Bodo Moeller]
1329 *) Fix to uni2asc() to cope with zero length Unicode strings.
1330 These are present in some PKCS#12 files.
1333 +) Finish off removing the remaining LHASH function pointer casts.
1334 There should no longer be any prototype-casting required when using
1335 the LHASH abstraction, and any casts that remain are "bugs". See
1336 the callback types and macros at the head of lhash.h for details
1337 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
1340 +) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
1341 If /dev/[u]random devices are not available or do not return enough
1342 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
1344 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
1345 /etc/entropy will be queried once each in this sequence, quering stops
1346 when enough entropy was collected without querying more sockets.
1349 +) Change the Unix RAND_poll() variant to be able to poll several
1350 random devices, as specified by DEVRANDOM, until a sufficient amount
1351 of data has been collected. We spend at most 10 ms on each file
1352 (select timeout) and read in non-blocking mode. DEVRANDOM now
1353 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
1354 (previously it was just the string "/dev/urandom"), so on typical
1355 platforms the 10 ms delay will never occur.
1356 Also separate out the Unix variant to its own file, rand_unix.c.
1357 For VMS, there's a currently-empty rand_vms.c.
1360 +) Move OCSP client related routines to ocsp_cl.c. These
1361 provide utility functions which an application needing
1362 to issue a request to an OCSP responder and analyse the
1363 response will typically need: as opposed to those which an
1364 OCSP responder itself would need which will be added later.
1366 OCSP_request_sign() signs an OCSP request with an API similar
1367 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
1368 response. OCSP_response_get1_basic() extracts basic response
1369 from response. OCSP_resp_find_status(): finds and extracts status
1370 information from an OCSP_CERTID structure (which will be created
1371 when the request structure is built). These are built from lower
1372 level functions which work on OCSP_SINGLERESP structures but
1373 wont normally be used unless the application wishes to examine
1374 extensions in the OCSP response for example.
1376 Replace nonce routines with a pair of functions.
1377 OCSP_request_add1_nonce() adds a nonce value and optionally
1378 generates a random value. OCSP_check_nonce() checks the
1379 validity of the nonce in an OCSP response.
1382 +) Change function OCSP_request_add() to OCSP_request_add0_id().
1383 This doesn't copy the supplied OCSP_CERTID and avoids the
1384 need to free up the newly created id. Change return type
1385 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
1386 This can then be used to add extensions to the request.
1387 Deleted OCSP_request_new(), since most of its functionality
1388 is now in OCSP_REQUEST_new() (and the case insensitive name
1389 clash) apart from the ability to set the request name which
1390 will be added elsewhere.
1393 +) Update OCSP API. Remove obsolete extensions argument from
1394 various functions. Extensions are now handled using the new
1395 OCSP extension code. New simple OCSP HTTP function which
1396 can be used to send requests and parse the response.
1399 +) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
1400 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
1401 uses the special reorder version of SET OF to sort the attributes
1402 and reorder them to match the encoded order. This resolves a long
1403 standing problem: a verify on a PKCS7 structure just after signing
1404 it used to fail because the attribute order did not match the
1405 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
1406 it uses the received order. This is necessary to tolerate some broken
1407 software that does not order SET OF. This is handled by encoding
1408 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
1409 to produce the required SET OF.
1412 +) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
1413 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
1414 files to get correct declarations of the ASN.1 item variables.
1417 +) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
1418 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
1419 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
1420 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
1421 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
1422 ASN1_ITEM and no wrapper functions.
1425 +) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
1426 replace the old function pointer based I/O routines. Change most of
1427 the *_d2i_bio() and *_d2i_fp() functions to use these.
1430 +) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
1431 lines, recognice more "algorithms" that can be deselected, and make
1432 it complain about algorithm deselection that isn't recognised.
1435 +) New ASN1 functions to handle dup, sign, verify, digest, pack and
1436 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
1437 to use new functions. Add NO_ASN1_OLD which can be set to remove
1438 some old style ASN1 functions: this can be used to determine if old
1439 code will still work when these eventually go away.
1442 +) New extension functions for OCSP structures, these follow the
1443 same conventions as certificates and CRLs.
1446 +) New function X509V3_add1_i2d(). This automatically encodes and
1447 adds an extension. Its behaviour can be customised with various
1448 flags to append, replace or delete. Various wrappers added for
1449 certifcates and CRLs.
1452 +) Fix to avoid calling the underlying ASN1 print routine when
1453 an extension cannot be parsed. Correct a typo in the
1454 OCSP_SERVICELOC extension. Tidy up print OCSP format.
1457 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
1458 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
1459 when writing a 32767 byte record.
1460 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
1462 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
1463 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
1465 (RSA objects have a reference count access to which is protected
1466 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
1467 so they are meant to be shared between threads.)
1468 [Bodo Moeller, Geoff Thorpe; original patch submitted by
1469 "Reddie, Steven" <Steven.Reddie@ca.com>]
1471 +) Make mkdef.pl parse some of the ASN1 macros and add apropriate
1472 entries for variables.
1475 *) Fix a deadlock in CRYPTO_mem_leaks().
1478 +) Add functionality to apps/openssl.c for detecting locking
1479 problems: As the program is single-threaded, all we have
1480 to do is register a locking callback using an array for
1481 storing which locks are currently held by the program.
1484 +) Use a lock around the call to CRYPTO_get_ex_new_index() in
1485 SSL_get_ex_data_X509_STORE_idx(), which is used in
1486 ssl_verify_cert_chain() and thus can be called at any time
1487 during TLS/SSL handshakes so that thread-safety is essential.
1488 Unfortunately, the ex_data design is not at all suited
1489 for multi-threaded use, so it probably should be abolished.
1492 +) Added Broadcom "ubsec" ENGINE to OpenSSL.
1493 [Broadcom, tweaked and integrated by Geoff Thorpe]
1495 +) Move common extension printing code to new function
1496 X509V3_print_extensions(). Reorganise OCSP print routines and
1497 implement some needed OCSP ASN1 functions. Add OCSP extensions.
1500 +) New function X509_signature_print() to remove duplication in some
1504 +) Add a special meaning when SET OF and SEQUENCE OF flags are both
1505 set (this was treated exactly the same as SET OF previously). This
1506 is used to reorder the STACK representing the structure to match the
1507 encoding. This will be used to get round a problem where a PKCS7
1508 structure which was signed could not be verified because the STACK
1509 order did not reflect the encoded order.
1512 +) Reimplement the OCSP ASN1 module using the new code.
1515 +) Update the X509V3 code to permit the use of an ASN1_ITEM structure
1516 for its ASN1 operations. The old style function pointers still exist
1517 for now but they will eventually go away.
1520 +) Merge in replacement ASN1 code from the ASN1 branch. This almost
1521 completely replaces the old ASN1 functionality with a table driven
1522 encoder and decoder which interprets an ASN1_ITEM structure describing
1523 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
1524 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
1525 has also been converted to the new form.
1528 +) Change BN_mod_exp_recp so that negative moduli are tolerated
1529 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
1530 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
1531 for negative moduli.
1534 +) Fix BN_uadd and BN_usub: Always return non-negative results instead
1535 of not touching the result's sign bit.
1538 +) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
1542 +) Changed the LHASH code to use prototypes for callbacks, and created
1543 macros to declare and implement thin (optionally static) functions
1544 that provide type-safety and avoid function pointer casting for the
1545 type-specific callbacks.
1548 *) Use better test patterns in bntest.
1551 +) Added Kerberos Cipher Suites to be used with TLS, as written in
1553 [Veers Staats <staatsvr@asc.hpc.mil>,
1554 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
1556 *) rand_win.c fix for Borland C.
1559 *) BN_rshift bugfix for n == 0.
1562 +) Reformat the FAQ so the different questions and answers can be divided
1563 in sections depending on the subject.
1566 +) Have the zlib compression code load ZLIB.DLL dynamically under
1570 +) New function BN_mod_sqrt for computing square roots modulo a prime
1571 (using the probabilistic Tonelli-Shanks algorithm unless
1572 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
1573 be handled deterministically).
1574 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
1576 *) Add a 'bctest' script that checks for some known 'bc' bugs
1577 so that 'make test' does not abort just because 'bc' is broken.
1580 *) Store verify_result within SSL_SESSION also for client side to
1581 avoid potential security hole. (Re-used sessions on the client side
1582 always resulted in verify_result==X509_V_OK, not using the original
1583 result of the server certificate verification.)
1586 +) Make BN_mod_inverse faster by explicitly handling small quotients
1587 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
1588 512 bits], about 30% for larger ones [1024 or 2048 bits].)
1591 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
1592 SSL3_RT_APPLICATION_DATA, return 0.
1593 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
1597 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
1598 releases, have been re-implemented by renaming the previous
1599 implementations of ssl2_read and ssl3_read to ssl2_read_internal
1600 and ssl3_read_internal, respectively, and adding 'peek' parameters
1601 to them. The new ssl[23]_{read,peek} functions are calls to
1602 ssl[23]_read_internal with the 'peek' flag set appropriately.
1603 A 'peek' parameter has also been added to ssl3_read_bytes, which
1604 does the actual work for ssl3_read_internal.
1607 +) New function BN_kronecker.
1610 +) Fix BN_gcd so that it works on negative inputs; the result is
1611 positive unless both parameters are zero.
1612 Previously something reasonably close to an infinite loop was
1613 possible because numbers could be growing instead of shrinking
1614 in the implementation of Euclid's algorithm.
1617 +) Fix BN_is_word() and BN_is_one() macros to take into account the
1618 sign of the number in question.
1620 Fix BN_is_word(a,w) to work correctly for w == 0.
1622 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
1623 because its test if the absolute value of 'a' equals 'w'.
1624 Note that BN_abs_is_word does *not* handle w == 0 reliably;
1625 it exists mostly for use in the implementations of BN_is_zero(),
1626 BN_is_one(), and BN_is_word().
1629 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
1630 the method-specific "init()" handler. Also clean up ex_data after
1631 calling the method-specific "finish()" handler. Previously, this was
1632 happening the other way round.
1635 +) New function BN_swap.
1638 +) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
1639 the exponentiation functions are more likely to produce reasonable
1640 results on negative inputs.
1643 +) Change BN_mod_mul so that the result is always non-negative.
1644 Previously, it could be negative if one of the factors was negative;
1645 I don't think anyone really wanted that behaviour.
1648 +) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
1649 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
1650 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
1651 and add new functions:
1660 BN_mod_lshift1_quick
1664 These functions always generate non-negative results.
1666 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
1667 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
1669 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
1670 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
1671 be reduced modulo m.
1672 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
1674 +) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
1675 was actually never needed) and in BN_mul(). The removal in BN_mul()
1676 required a small change in bn_mul_part_recursive() and the addition
1677 of the functions bn_cmp_part_words(), bn_sub_part_words() and
1678 bn_add_part_words(), which do the same thing as bn_cmp_words(),
1679 bn_sub_words() and bn_add_words() except they take arrays with
1683 +) In 'openssl passwd', verify passwords read from the terminal
1684 unless the '-salt' option is used (which usually means that
1685 verification would just waste user's time since the resulting
1686 hash is going to be compared with some given password hash)
1687 or the new '-noverify' option is used.
1689 This is an incompatible change, but it does not affect
1690 non-interactive use of 'openssl passwd' (passwords on the command
1691 line, '-stdin' option, '-in ...' option) and thus should not
1695 +) Remove all references to RSAref, since there's no more need for it.
1698 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
1699 The previous value, 12, was not always sufficient for BN_mod_exp().
1702 +) Make DSO load along a path given through an environment variable
1703 (SHLIB_PATH) with shl_load().
1706 +) Constify the ENGINE code as a result of BIGNUM constification.
1707 Also constify the RSA code and most things related to it. In a
1708 few places, most notable in the depth of the ASN.1 code, ugly
1709 casts back to non-const were required (to be solved at a later
1713 +) Make it so the openssl application has all engines loaded by default.
1716 +) Constify the BIGNUM routines a little more.
1719 *) Make sure that shared libraries get the internal name engine with
1720 the full version number and not just 0. This should mark the
1721 shared libraries as not backward compatible. Of course, this should
1722 be changed again when we can guarantee backward binary compatibility.
1725 +) Add the following functions:
1727 ENGINE_load_cswift()
1729 ENGINE_load_atalla()
1731 ENGINE_load_builtin_engines()
1733 That way, an application can itself choose if external engines that
1734 are built-in in OpenSSL shall ever be used or not. The benefit is
1735 that applications won't have to be linked with libdl or other dso
1736 libraries unless it's really needed.
1738 Changed 'openssl engine' to load all engines on demand.
1739 Changed the engine header files to avoid the duplication of some
1740 declarations (they differed!).
1743 +) 'openssl engine' can now list capabilities.
1746 +) Better error reporting in 'openssl engine'.
1749 +) Never call load_dh_param(NULL) in s_server.
1752 +) Add engine application. It can currently list engines by name and
1753 identity, and test if they are actually available.
1756 +) Add support for shared libraries under Irix.
1757 [Albert Chin-A-Young <china@thewrittenword.com>]
1759 +) Improve RPM specification file by forcing symbolic linking and making
1760 sure the installed documentation is also owned by root.root.
1761 [Damien Miller <djm@mindrot.org>]
1763 +) Add configuration option to build on Linux on both big-endian and
1765 [Ralf Baechle <ralf@uni-koblenz.de>]
1767 +) Give the OpenSSL applications more possibilities to make use of
1768 keys (public as well as private) handled by engines.
1771 +) Add OCSP code that comes from CertCo.
1774 +) Add VMS support for the Rijndael code.
1777 +) Added untested support for Nuron crypto accelerator.
1780 +) Add support for external cryptographic devices. This code was
1781 previously distributed separately as the "engine" branch.
1782 [Geoff Thorpe, Richard Levitte]
1784 +) Rework the filename-translation in the DSO code. It is now possible to
1785 have far greater control over how a "name" is turned into a filename
1786 depending on the operating environment and any oddities about the
1787 different shared library filenames on each system.
1790 +) Support threads on FreeBSD-elf in Configure.
1793 +) Add the possibility to create shared libraries on HP-UX
1796 +) Fix for SHA1 assembly problem with MASM: it produces
1797 warnings about corrupt line number information when assembling
1798 with debugging information. This is caused by the overlapping
1800 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
1803 NCONF_get_number() has no error checking at all. As a replacement,
1804 NCONF_get_number_e() is defined (_e for "error checking") and is
1805 promoted strongly. The old NCONF_get_number is kept around for
1806 binary backward compatibility.
1807 Make it possible for methods to load from something other than a BIO,
1808 by providing a function pointer that is given a name instead of a BIO.
1809 For example, this could be used to load configuration data from an
1813 *) Fix typo in get_cert_by_subject() in by_dir.c
1814 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
1816 *) Rework the system to generate shared libraries:
1818 - Make note of the expected extension for the shared libraries and
1819 if there is a need for symbolic links from for example libcrypto.so.0
1820 to libcrypto.so.0.9.7. There is extended info in Configure for
1823 - Make as few rebuilds of the shared libraries as possible.
1825 - Still avoid linking the OpenSSL programs with the shared libraries.
1827 - When installing, install the shared libraries separately from the
1831 +) Fix for non blocking accept BIOs. Added new I/O special reason
1832 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
1833 with non blocking I/O was not possible because no retry code was
1834 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
1838 +) Added the beginnings of Rijndael support.
1841 +) Fix for bug in DirectoryString mask setting. Add support for
1842 X509_NAME_print_ex() in 'req' and X509_print_ex() function
1843 to allow certificate printing to more controllable, additional
1844 'certopt' option to 'x509' to allow new printing options to be
1848 +) Clean old EAY MD5 hack from e_os.h.
1851 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
1853 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
1854 and not in SSL_clear because the latter is also used by the
1855 accept/connect functions; previously, the settings made by
1856 SSL_set_read_ahead would be lost during the handshake.
1857 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
1859 *) Correct util/mkdef.pl to be selective about disabled algorithms.
1860 Previously, it would create entries for disableed algorithms no
1864 *) Added several new manual pages for SSL_* function.
1867 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
1869 *) In ssl23_get_client_hello, generate an error message when faced
1870 with an initial SSL 3.0/TLS record that is too small to contain the
1871 first two bytes of the ClientHello message, i.e. client_version.
1872 (Note that this is a pathologic case that probably has never happened
1873 in real life.) The previous approach was to use the version number
1874 from the record header as a substitute; but our protocol choice
1875 should not depend on that one because it is not authenticated
1876 by the Finished messages.
1879 *) More robust randomness gathering functions for Windows.
1880 [Jeffrey Altman <jaltman@columbia.edu>]
1882 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
1883 not set then we don't setup the error code for issuer check errors
1884 to avoid possibly overwriting other errors which the callback does
1885 handle. If an application does set the flag then we assume it knows
1886 what it is doing and can handle the new informational codes
1890 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
1891 a general "ANY" type, as such it should be able to decode anything
1892 including tagged types. However it didn't check the class so it would
1893 wrongly interpret tagged types in the same way as their universal
1894 counterpart and unknown types were just rejected. Changed so that the
1895 tagged and unknown types are handled in the same way as a SEQUENCE:
1896 that is the encoding is stored intact. There is also a new type
1897 "V_ASN1_OTHER" which is used when the class is not universal, in this
1898 case we have no idea what the actual type is so we just lump them all
1902 *) On VMS, stdout may very well lead to a file that is written to
1903 in a record-oriented fashion. That means that every write() will
1904 write a separate record, which will be read separately by the
1905 programs trying to read from it. This can be very confusing.
1907 The solution is to put a BIO filter in the way that will buffer
1908 text until a linefeed is reached, and then write everything a
1909 line at a time, so every record written will be an actual line,
1910 not chunks of lines and not (usually doesn't happen, but I've
1911 seen it once) several lines in one record. BIO_f_linebuffer() is
1914 Currently, it's a VMS-only method, because that's where it has
1915 been tested well enough.
1918 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
1919 it can return incorrect results.
1920 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
1921 but it was in 0.9.6-beta[12].)
1924 *) Disable the check for content being present when verifying detached
1925 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
1926 include zero length content when signing messages.
1929 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
1930 BIO_ctrl (for BIO pairs).
1933 *) Add DSO method for VMS.
1936 *) Bug fix: Montgomery multiplication could produce results with the
1940 *) Add RPM specification openssl.spec and modify it to build three
1941 packages. The default package contains applications, application
1942 documentation and run-time libraries. The devel package contains
1943 include files, static libraries and function documentation. The
1944 doc package contains the contents of the doc directory. The original
1945 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
1948 *) Add a large number of documentation files for many SSL routines.
1949 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
1951 *) Add a configuration entry for Sony News 4.
1952 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
1954 *) Don't set the two most significant bits to one when generating a
1955 random number < q in the DSA library.
1958 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
1959 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
1960 the underlying transport is blocking) if a handshake took place.
1961 (The default behaviour is needed by applications such as s_client
1962 and s_server that use select() to determine when to use SSL_read;
1963 but for applications that know in advance when to expect data, it
1964 just makes things more complicated.)
1967 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
1971 *) Add a few more EBCDIC conditionals that make `req' and `x509'
1972 work better on such systems.
1973 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
1975 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
1976 Update PKCS12_parse() so it copies the friendlyName and the
1977 keyid to the certificates aux info.
1980 *) Fix bug in PKCS7_verify() which caused an infinite loop
1981 if there was more than one signature.
1982 [Sven Uszpelkat <su@celocom.de>]
1984 *) Major change in util/mkdef.pl to include extra information
1985 about each symbol, as well as presentig variables as well
1986 as functions. This change means that there's n more need
1987 to rebuild the .num files when some algorithms are excluded.
1990 *) Allow the verify time to be set by an application,
1991 rather than always using the current time.
1994 *) Phase 2 verify code reorganisation. The certificate
1995 verify code now looks up an issuer certificate by a
1996 number of criteria: subject name, authority key id
1997 and key usage. It also verifies self signed certificates
1998 by the same criteria. The main comparison function is
1999 X509_check_issued() which performs these checks.
2001 Lot of changes were necessary in order to support this
2002 without completely rewriting the lookup code.
2004 Authority and subject key identifier are now cached.
2006 The LHASH 'certs' is X509_STORE has now been replaced
2007 by a STACK_OF(X509_OBJECT). This is mainly because an
2008 LHASH can't store or retrieve multiple objects with
2009 the same hash value.
2011 As a result various functions (which were all internal
2012 use only) have changed to handle the new X509_STORE
2013 structure. This will break anything that messed round
2014 with X509_STORE internally.
2016 The functions X509_STORE_add_cert() now checks for an
2017 exact match, rather than just subject name.
2019 The X509_STORE API doesn't directly support the retrieval
2020 of multiple certificates matching a given criteria, however
2021 this can be worked round by performing a lookup first
2022 (which will fill the cache with candidate certificates)
2023 and then examining the cache for matches. This is probably
2024 the best we can do without throwing out X509_LOOKUP
2025 entirely (maybe later...).
2027 The X509_VERIFY_CTX structure has been enhanced considerably.
2029 All certificate lookup operations now go via a get_issuer()
2030 callback. Although this currently uses an X509_STORE it
2031 can be replaced by custom lookups. This is a simple way
2032 to bypass the X509_STORE hackery necessary to make this
2033 work and makes it possible to use more efficient techniques
2034 in future. A very simple version which uses a simple
2035 STACK for its trusted certificate store is also provided
2036 using X509_STORE_CTX_trusted_stack().
2038 The verify_cb() and verify() callbacks now have equivalents
2039 in the X509_STORE_CTX structure.
2041 X509_STORE_CTX also has a 'flags' field which can be used
2042 to customise the verify behaviour.
2045 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
2046 excludes S/MIME capabilities.
2049 *) When a certificate request is read in keep a copy of the
2050 original encoding of the signed data and use it when outputing
2051 again. Signatures then use the original encoding rather than
2052 a decoded, encoded version which may cause problems if the
2053 request is improperly encoded.
2056 *) For consistency with other BIO_puts implementations, call
2057 buffer_write(b, ...) directly in buffer_puts instead of calling
2060 In BIO_puts, increment b->num_write as in BIO_write.
2061 [Peter.Sylvester@EdelWeb.fr]
2063 *) Fix BN_mul_word for the case where the word is 0. (We have to use
2064 BN_zero, we may not return a BIGNUM with an array consisting of
2068 *) Avoid calling abort() from within the library when problems are
2069 detected, except if preprocessor symbols have been defined
2070 (such as REF_CHECK, BN_DEBUG etc.).
2073 *) New openssl application 'rsautl'. This utility can be
2074 used for low level RSA operations. DER public key
2075 BIO/fp routines also added.
2078 *) New Configure entry and patches for compiling on QNX 4.
2079 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
2081 *) A demo state-machine implementation was sponsored by
2082 Nuron (http://www.nuron.com/) and is now available in
2083 demos/state_machine.
2086 *) New options added to the 'dgst' utility for signature
2087 generation and verification.
2090 *) Unrecognized PKCS#7 content types are now handled via a
2091 catch all ASN1_TYPE structure. This allows unsupported
2092 types to be stored as a "blob" and an application can
2093 encode and decode it manually.
2096 *) Fix various signed/unsigned issues to make a_strex.c
2098 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
2100 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
2101 length if passed a buffer. ASN1_INTEGER_to_BN failed
2102 if passed a NULL BN and its argument was negative.
2103 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
2105 *) Modification to PKCS#7 encoding routines to output definite
2106 length encoding. Since currently the whole structures are in
2107 memory there's not real point in using indefinite length
2108 constructed encoding. However if OpenSSL is compiled with
2109 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
2112 *) Added BIO_vprintf() and BIO_vsnprintf().
2115 *) Added more prefixes to parse for in the the strings written
2116 through a logging bio, to cover all the levels that are available
2117 through syslog. The prefixes are now:
2119 PANIC, EMERG, EMR => LOG_EMERG
2120 ALERT, ALR => LOG_ALERT
2121 CRIT, CRI => LOG_CRIT
2122 ERROR, ERR => LOG_ERR
2123 WARNING, WARN, WAR => LOG_WARNING
2124 NOTICE, NOTE, NOT => LOG_NOTICE
2125 INFO, INF => LOG_INFO
2126 DEBUG, DBG => LOG_DEBUG
2128 and as before, if none of those prefixes are present at the
2129 beginning of the string, LOG_ERR is chosen.
2131 On Win32, the LOG_* levels are mapped according to this:
2133 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
2134 LOG_WARNING => EVENTLOG_WARNING_TYPE
2135 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
2139 *) Made it possible to reconfigure with just the configuration
2140 argument "reconf" or "reconfigure". The command line arguments
2141 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
2142 and are retrieved from there when reconfiguring.
2146 [Assar Westerlund <assar@sics.se>, Richard Levitte]
2148 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
2151 *) The obj_dat.pl script was messing up the sorting of object
2152 names. The reason was that it compared the quoted version
2153 of strings as a result "OCSP" > "OCSP Signing" because
2154 " > SPACE. Changed script to store unquoted versions of
2155 names and add quotes on output. It was also omitting some
2156 names from the lookup table if they were given a default
2157 value (that is if SN is missing it is given the same
2158 value as LN and vice versa), these are now added on the
2159 grounds that if an object has a name we should be able to
2160 look it up. Finally added warning output when duplicate
2161 short or long names are found.
2164 *) Changes needed for Tandem NSK.
2165 [Scott Uroff <scott@xypro.com>]
2167 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
2168 RSA_padding_check_SSLv23(), special padding was never detected
2169 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
2170 version rollback attacks was not effective.
2172 In s23_clnt.c, don't use special rollback-attack detection padding
2173 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
2174 client; similarly, in s23_srvr.c, don't do the rollback check if
2175 SSL 2.0 is the only protocol enabled in the server.
2178 *) Make it possible to get hexdumps of unprintable data with 'openssl
2179 asn1parse'. By implication, the functions ASN1_parse_dump() and
2180 BIO_dump_indent() are added.
2183 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
2184 these print out strings and name structures based on various
2185 flags including RFC2253 support and proper handling of
2186 multibyte characters. Added options to the 'x509' utility
2187 to allow the various flags to be set.
2190 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
2191 Also change the functions X509_cmp_current_time() and
2192 X509_gmtime_adj() work with an ASN1_TIME structure,
2193 this will enable certificates using GeneralizedTime in validity
2194 dates to be checked.
2197 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
2198 negative public key encodings) on by default,
2199 NO_NEG_PUBKEY_BUG can be set to disable it.
2202 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
2203 content octets. An i2c_ASN1_OBJECT is unnecessary because
2204 the encoding can be trivially obtained from the structure.
2207 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
2208 not read locks (CRYPTO_r_[un]lock).
2211 *) A first attempt at creating official support for shared
2212 libraries through configuration. I've kept it so the
2213 default is static libraries only, and the OpenSSL programs
2214 are always statically linked for now, but there are
2215 preparations for dynamic linking in place.
2216 This has been tested on Linux and Tru64.
2219 *) Randomness polling function for Win9x, as described in:
2220 Peter Gutmann, Software Generation of Practically Strong
2224 *) Fix so PRNG is seeded in req if using an already existing
2228 *) New options to smime application. -inform and -outform
2229 allow alternative formats for the S/MIME message including
2230 PEM and DER. The -content option allows the content to be
2231 specified separately. This should allow things like Netscape
2232 form signing output easier to verify.
2235 *) Fix the ASN1 encoding of tags using the 'long form'.
2238 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
2239 STRING types. These convert content octets to and from the
2240 underlying type. The actual tag and length octets are
2241 already assumed to have been read in and checked. These
2242 are needed because all other string types have virtually
2243 identical handling apart from the tag. By having versions
2244 of the ASN1 functions that just operate on content octets
2245 IMPLICIT tagging can be handled properly. It also allows
2246 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
2247 and ASN1_INTEGER are identical apart from the tag.
2250 *) Change the handling of OID objects as follows:
2252 - New object identifiers are inserted in objects.txt, following
2253 the syntax given in objects.README.
2254 - objects.pl is used to process obj_mac.num and create a new
2256 - obj_dat.pl is used to create a new obj_dat.h, using the data in
2259 This is currently kind of a hack, and the perl code in objects.pl
2260 isn't very elegant, but it works as I intended. The simplest way
2261 to check that it worked correctly is to look in obj_dat.h and
2262 check the array nid_objs and make sure the objects haven't moved
2263 around (this is important!). Additions are OK, as well as
2264 consistent name changes.
2267 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
2270 *) Addition of the command line parameter '-rand file' to 'openssl req'.
2271 The given file adds to whatever has already been seeded into the
2272 random pool through the RANDFILE configuration file option or
2273 environment variable, or the default random state file.
2276 *) mkstack.pl now sorts each macro group into lexical order.
2277 Previously the output order depended on the order the files
2278 appeared in the directory, resulting in needless rewriting
2282 *) Patches to make OpenSSL compile under Win32 again. Mostly
2283 work arounds for the VC++ problem that it treats func() as
2284 func(void). Also stripped out the parts of mkdef.pl that
2285 added extra typesafe functions: these no longer exist.
2288 *) Reorganisation of the stack code. The macros are now all
2289 collected in safestack.h . Each macro is defined in terms of
2290 a "stack macro" of the form SKM_<name>(type, a, b). The
2291 DEBUG_SAFESTACK is now handled in terms of function casts,
2292 this has the advantage of retaining type safety without the
2293 use of additional functions. If DEBUG_SAFESTACK is not defined
2294 then the non typesafe macros are used instead. Also modified the
2295 mkstack.pl script to handle the new form. Needs testing to see
2296 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
2297 the default if no major problems. Similar behaviour for ASN1_SET_OF
2298 and PKCS12_STACK_OF.
2301 *) When some versions of IIS use the 'NET' form of private key the
2302 key derivation algorithm is different. Normally MD5(password) is
2303 used as a 128 bit RC4 key. In the modified case
2304 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
2305 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
2306 as the old Netscape_RSA functions except they have an additional
2307 'sgckey' parameter which uses the modified algorithm. Also added
2308 an -sgckey command line option to the rsa utility. Thanks to
2309 Adrian Peck <bertie@ncipher.com> for posting details of the modified
2310 algorithm to openssl-dev.
2313 *) The evp_local.h macros were using 'c.##kname' which resulted in
2314 invalid expansion on some systems (SCO 5.0.5 for example).
2315 Corrected to 'c.kname'.
2316 [Phillip Porch <root@theporch.com>]
2318 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
2319 a STACK of email addresses from a certificate or request, these look
2320 in the subject name and the subject alternative name extensions and
2321 omit any duplicate addresses.
2324 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
2325 This makes DSA verification about 2 % faster.
2328 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
2329 (meaning that now 2^5 values will be precomputed, which is only 4 KB
2330 plus overhead for 1024 bit moduli).
2331 This makes exponentiations about 0.5 % faster for 1024 bit
2332 exponents (as measured by "openssl speed rsa2048").
2335 *) Rename memory handling macros to avoid conflicts with other
2337 Malloc => OPENSSL_malloc
2338 Malloc_locked => OPENSSL_malloc_locked
2339 Realloc => OPENSSL_realloc
2340 Free => OPENSSL_free
2343 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
2344 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
2347 *) CygWin32 support.
2348 [John Jarvie <jjarvie@newsguy.com>]
2350 *) The type-safe stack code has been rejigged. It is now only compiled
2351 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
2352 by default all type-specific stack functions are "#define"d back to
2353 standard stack functions. This results in more streamlined output
2354 but retains the type-safety checking possibilities of the original
2358 *) The STACK code has been cleaned up, and certain type declarations
2359 that didn't make a lot of sense have been brought in line. This has
2360 also involved a cleanup of sorts in safestack.h to more correctly
2361 map type-safe stack functions onto their plain stack counterparts.
2362 This work has also resulted in a variety of "const"ifications of
2363 lots of the code, especially "_cmp" operations which should normally
2364 be prototyped with "const" parameters anyway.
2367 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
2368 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
2369 (The PRNG state consists of two parts, the large pool 'state' and 'md',
2370 where all of 'md' is used each time the PRNG is used, but 'state'
2371 is used only indexed by a cyclic counter. As entropy may not be
2372 well distributed from the beginning, 'md' is important as a
2373 chaining variable. However, the output function chains only half
2374 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
2375 all of 'md', and seeding with STATE_SIZE dummy bytes will result
2376 in all of 'state' being rewritten, with the new values depending
2377 on virtually all of 'md'. This overcomes the 80 bit limitation.)
2380 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
2381 the handshake is continued after ssl_verify_cert_chain();
2382 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
2383 can lead to 'unexplainable' connection aborts later.
2384 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
2386 *) Major EVP API cipher revision.
2387 Add hooks for extra EVP features. This allows various cipher
2388 parameters to be set in the EVP interface. Support added for variable
2389 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
2390 setting of RC2 and RC5 parameters.
2392 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
2395 Remove lots of duplicated code from the EVP library. For example *every*
2396 cipher init() function handles the 'iv' in the same way according to the
2397 cipher mode. They also all do nothing if the 'key' parameter is NULL and
2398 for CFB and OFB modes they zero ctx->num.
2400 New functionality allows removal of S/MIME code RC2 hack.
2402 Most of the routines have the same form and so can be declared in terms
2405 By shifting this to the top level EVP_CipherInit() it can be removed from
2406 all individual ciphers. If the cipher wants to handle IVs or keys
2407 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
2410 Change lots of functions like EVP_EncryptUpdate() to now return a
2411 value: although software versions of the algorithms cannot fail
2412 any installed hardware versions can.
2415 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
2416 this option is set, tolerate broken clients that send the negotiated
2417 protocol version number instead of the requested protocol version
2421 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
2422 i.e. non-zero for export ciphersuites, zero otherwise.
2423 Previous versions had this flag inverted, inconsistent with
2424 rsa_tmp_cb (..._TMP_RSA_CB).
2425 [Bodo Moeller; problem reported by Amit Chopra]
2427 *) Add missing DSA library text string. Work around for some IIS
2428 key files with invalid SEQUENCE encoding.
2431 *) Add a document (doc/standards.txt) that list all kinds of standards
2432 and so on that are implemented in OpenSSL.
2435 *) Enhance c_rehash script. Old version would mishandle certificates
2436 with the same subject name hash and wouldn't handle CRLs at all.
2437 Added -fingerprint option to crl utility, to support new c_rehash
2441 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
2444 *) Fix for SSL server purpose checking. Server checking was
2445 rejecting certificates which had extended key usage present
2446 but no ssl client purpose.
2447 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
2449 *) Make PKCS#12 code work with no password. The PKCS#12 spec
2450 is a little unclear about how a blank password is handled.
2451 Since the password in encoded as a BMPString with terminating
2452 double NULL a zero length password would end up as just the
2453 double NULL. However no password at all is different and is
2454 handled differently in the PKCS#12 key generation code. NS
2455 treats a blank password as zero length. MSIE treats it as no
2456 password on export: but it will try both on import. We now do
2457 the same: PKCS12_parse() tries zero length and no password if
2458 the password is set to "" or NULL (NULL is now a valid password:
2459 it wasn't before) as does the pkcs12 application.
2462 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
2463 perror when PEM_read_bio_X509_REQ fails, the error message must
2464 be obtained from the error queue.
2467 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
2468 it in ERR_remove_state if appropriate, and change ERR_get_state
2469 accordingly to avoid race conditions (this is necessary because
2470 thread_hash is no longer constant once set).
2473 *) Bugfix for linux-elf makefile.one.
2476 *) RSA_get_default_method() will now cause a default
2477 RSA_METHOD to be chosen if one doesn't exist already.
2478 Previously this was only set during a call to RSA_new()
2479 or RSA_new_method(NULL) meaning it was possible for
2480 RSA_get_default_method() to return NULL.
2483 *) Added native name translation to the existing DSO code
2484 that will convert (if the flag to do so is set) filenames
2485 that are sufficiently small and have no path information
2486 into a canonical native form. Eg. "blah" converted to
2487 "libblah.so" or "blah.dll" etc.
2490 *) New function ERR_error_string_n(e, buf, len) which is like
2491 ERR_error_string(e, buf), but writes at most 'len' bytes
2492 including the 0 terminator. For ERR_error_string_n, 'buf'
2494 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
2496 *) CONF library reworked to become more general. A new CONF
2497 configuration file reader "class" is implemented as well as a
2498 new functions (NCONF_*, for "New CONF") to handle it. The now
2499 old CONF_* functions are still there, but are reimplemented to
2500 work in terms of the new functions. Also, a set of functions
2501 to handle the internal storage of the configuration data is
2502 provided to make it easier to write new configuration file
2503 reader "classes" (I can definitely see something reading a
2504 configuration file in XML format, for example), called _CONF_*,
2505 or "the configuration storage API"...
2507 The new configuration file reading functions are:
2509 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
2510 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
2512 NCONF_default, NCONF_WIN32
2514 NCONF_dump_fp, NCONF_dump_bio
2516 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
2517 NCONF_new creates a new CONF object. This works in the same way
2518 as other interfaces in OpenSSL, like the BIO interface.
2519 NCONF_dump_* dump the internal storage of the configuration file,
2520 which is useful for debugging. All other functions take the same
2521 arguments as the old CONF_* functions wth the exception of the
2522 first that must be a `CONF *' instead of a `LHASH *'.
2524 To make it easer to use the new classes with the old CONF_* functions,
2525 the function CONF_set_default_method is provided.
2528 *) Add '-tls1' option to 'openssl ciphers', which was already
2529 mentioned in the documentation but had not been implemented.
2530 (This option is not yet really useful because even the additional
2531 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
2534 *) Initial DSO code added into libcrypto for letting OpenSSL (and
2535 OpenSSL-based applications) load shared libraries and bind to
2536 them in a portable way.
2537 [Geoff Thorpe, with contributions from Richard Levitte]
2539 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
2541 *) Make sure _lrotl and _lrotr are only used with MSVC.
2543 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
2544 (the default implementation of RAND_status).
2546 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
2547 to '-clrext' (= clear extensions), as intended and documented.
2548 [Bodo Moeller; inconsistency pointed out by Michael Attili
2551 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
2552 was larger than the MD block size.
2553 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
2555 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
2556 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
2557 using the passed key: if the passed key was a private key the result
2558 of X509_print(), for example, would be to print out all the private key
2562 *) des_quad_cksum() byte order bug fix.
2563 [Ulf Möller, using the problem description in krb4-0.9.7, where
2564 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
2566 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
2568 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
2570 *) For easily testing in shell scripts whether some command
2571 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
2572 returns with exit code 0 iff no command of the given name is available.
2573 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
2574 the output goes to stdout and nothing is printed to stderr.
2575 Additional arguments are always ignored.
2577 Since for each cipher there is a command of the same name,
2578 the 'no-cipher' compilation switches can be tested this way.
2580 ('openssl no-XXX' is not able to detect pseudo-commands such
2581 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
2584 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
2587 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
2588 is set; it will be thrown away anyway because each handshake creates
2590 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
2591 to parameters -- in previous versions (since OpenSSL 0.9.3) the
2592 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
2593 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
2596 *) New s_client option -ign_eof: EOF at stdin is ignored, and
2597 'Q' and 'R' lose their special meanings (quit/renegotiate).
2598 This is part of what -quiet does; unlike -quiet, -ign_eof
2599 does not suppress any output.
2602 *) Add compatibility options to the purpose and trust code. The
2603 purpose X509_PURPOSE_ANY is "any purpose" which automatically
2604 accepts a certificate or CA, this was the previous behaviour,
2605 with all the associated security issues.
2607 X509_TRUST_COMPAT is the old trust behaviour: only and
2608 automatically trust self signed roots in certificate store. A
2609 new trust setting X509_TRUST_DEFAULT is used to specify that
2610 a purpose has no associated trust setting and it should instead
2611 use the value in the default purpose.
2614 *) Fix the PKCS#8 DSA private key code so it decodes keys again
2615 and fix a memory leak.
2618 *) In util/mkerr.pl (which implements 'make errors'), preserve
2619 reason strings from the previous version of the .c file, as
2620 the default to have only downcase letters (and digits) in
2621 automatically generated reasons codes is not always appropriate.
2624 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
2625 using strerror. Previously, ERR_reason_error_string() returned
2626 library names as reason strings for SYSerr; but SYSerr is a special
2627 case where small numbers are errno values, not library numbers.
2630 *) Add '-dsaparam' option to 'openssl dhparam' application. This
2631 converts DSA parameters into DH parameters. (When creating parameters,
2632 DSA_generate_parameters is used.)
2635 *) Include 'length' (recommended exponent length) in C code generated
2636 by 'openssl dhparam -C'.
2639 *) The second argument to set_label in perlasm was already being used
2640 so couldn't be used as a "file scope" flag. Moved to third argument
2644 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
2645 instead of RAND_bytes for encryption IVs and salts.
2648 *) Include RAND_status() into RAND_METHOD instead of implementing
2649 it only for md_rand.c Otherwise replacing the PRNG by calling
2650 RAND_set_rand_method would be impossible.
2653 *) Don't let DSA_generate_key() enter an infinite loop if the random
2654 number generation fails.
2657 *) New 'rand' application for creating pseudo-random output.
2660 *) Added configuration support for Linux/IA64
2661 [Rolf Haberrecker <rolf@suse.de>]
2663 *) Assembler module support for Mingw32.
2666 *) Shared library support for HPUX (in shlib/).
2667 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
2669 *) Shared library support for Solaris gcc.
2670 [Lutz Behnke <behnke@trustcenter.de>]
2672 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
2674 *) PKCS7_encrypt() was adding text MIME headers twice because they
2675 were added manually and by SMIME_crlf_copy().
2678 *) In bntest.c don't call BN_rand with zero bits argument.
2679 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
2681 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
2682 case was implemented. This caused BN_div_recp() to fail occasionally.
2685 *) Add an optional second argument to the set_label() in the perl
2686 assembly language builder. If this argument exists and is set
2687 to 1 it signals that the assembler should use a symbol whose
2688 scope is the entire file, not just the current function. This
2689 is needed with MASM which uses the format label:: for this scope.
2690 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
2692 *) Change the ASN1 types so they are typedefs by default. Before
2693 almost all types were #define'd to ASN1_STRING which was causing
2694 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
2698 *) Change names of new functions to the new get1/get0 naming
2699 convention: After 'get1', the caller owns a reference count
2700 and has to call ..._free; 'get0' returns a pointer to some
2701 data structure without incrementing reference counters.
2702 (Some of the existing 'get' functions increment a reference
2703 counter, some don't.)
2704 Similarly, 'set1' and 'add1' functions increase reference
2705 counters or duplicate objects.
2708 *) Allow for the possibility of temp RSA key generation failure:
2709 the code used to assume it always worked and crashed on failure.
2712 *) Fix potential buffer overrun problem in BIO_printf().
2713 [Ulf Möller, using public domain code by Patrick Powell; problem
2714 pointed out by David Sacerdote <das33@cornell.edu>]
2716 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
2717 RAND_egd() and RAND_status(). In the command line application,
2718 the EGD socket can be specified like a seed file using RANDFILE
2722 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
2723 Some CAs (e.g. Verisign) distribute certificates in this form.
2726 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
2727 list to exclude them. This means that no special compilation option
2728 is needed to use anonymous DH: it just needs to be included in the
2732 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
2733 EVP_MD_type. The old functionality is available in a new macro called
2734 EVP_MD_md(). Change code that uses it and update docs.
2737 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
2738 where the 'void *' argument is replaced by a function pointer argument.
2739 Previously 'void *' was abused to point to functions, which works on
2740 many platforms, but is not correct. As these functions are usually
2741 called by macros defined in OpenSSL header files, most source code
2742 should work without changes.
2745 *) <openssl/opensslconf.h> (which is created by Configure) now contains
2746 sections with information on -D... compiler switches used for
2747 compiling the library so that applications can see them. To enable
2748 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
2749 must be defined. E.g.,
2750 #define OPENSSL_ALGORITHM_DEFINES
2751 #include <openssl/opensslconf.h>
2752 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
2753 [Richard Levitte, Ulf and Bodo Möller]
2755 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
2759 *) Change the 'other' type in certificate aux info to a STACK_OF
2760 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
2761 the required ASN1 format: arbitrary types determined by an OID.
2764 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
2765 argument to 'req'. This is not because the function is newer or
2766 better than others it just uses the work 'NEW' in the certificate
2767 request header lines. Some software needs this.
2770 *) Reorganise password command line arguments: now passwords can be
2771 obtained from various sources. Delete the PEM_cb function and make
2772 it the default behaviour: i.e. if the callback is NULL and the
2773 usrdata argument is not NULL interpret it as a null terminated pass
2774 phrase. If usrdata and the callback are NULL then the pass phrase
2775 is prompted for as usual.
2778 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
2779 the support is automatically enabled. The resulting binaries will
2780 autodetect the card and use it if present.
2781 [Ben Laurie and Compaq Inc.]
2783 *) Work around for Netscape hang bug. This sends certificate request
2784 and server done in one record. Since this is perfectly legal in the
2785 SSL/TLS protocol it isn't a "bug" option and is on by default. See
2786 the bugs/SSLv3 entry for more info.
2789 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
2792 *) Add -rand argument to smime and pkcs12 applications and read/write
2796 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
2799 *) Add command line password options to the remaining applications.
2802 *) Bug fix for BN_div_recp() for numerators with an even number of
2806 *) More tests in bntest.c, and changed test_bn output.
2809 *) ./config recognizes MacOS X now.
2812 *) Bug fix for BN_div() when the first words of num and divsor are
2813 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
2816 *) Add support for various broken PKCS#8 formats, and command line
2817 options to produce them.
2820 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
2821 get temporary BIGNUMs from a BN_CTX.
2824 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
2828 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
2829 include a #define from the old name to the new. The original intent
2830 was that statically linked binaries could for example just call
2831 SSLeay_add_all_ciphers() to just add ciphers to the table and not
2832 link with digests. This never worked becayse SSLeay_add_all_digests()
2833 and SSLeay_add_all_ciphers() were in the same source file so calling
2834 one would link with the other. They are now in separate source files.
2837 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
2840 *) Use a less unusual form of the Miller-Rabin primality test (it used
2841 a binary algorithm for exponentiation integrated into the Miller-Rabin
2842 loop, our standard modexp algorithms are faster).
2845 *) Support for the EBCDIC character set completed.
2846 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
2848 *) Source code cleanups: use const where appropriate, eliminate casts,
2849 use void * instead of char * in lhash.
2852 *) Bugfix: ssl3_send_server_key_exchange was not restartable
2853 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
2854 this the server could overwrite ephemeral keys that the client
2858 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
2859 using 50 iterations of the Rabin-Miller test.
2861 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
2862 iterations of the Rabin-Miller test as required by the appendix
2863 to FIPS PUB 186[-1]) instead of DSA_is_prime.
2864 As BN_is_prime_fasttest includes trial division, DSA parameter
2865 generation becomes much faster.
2867 This implies a change for the callback functions in DSA_is_prime
2868 and DSA_generate_parameters: The callback function is called once
2869 for each positive witness in the Rabin-Miller test, not just
2870 occasionally in the inner loop; and the parameters to the
2871 callback function now provide an iteration count for the outer
2872 loop rather than for the current invocation of the inner loop.
2873 DSA_generate_parameters additionally can call the callback
2874 function with an 'iteration count' of -1, meaning that a
2875 candidate has passed the trial division test (when q is generated
2876 from an application-provided seed, trial division is skipped).
2879 *) New function BN_is_prime_fasttest that optionally does trial
2880 division before starting the Rabin-Miller test and has
2881 an additional BN_CTX * argument (whereas BN_is_prime always
2882 has to allocate at least one BN_CTX).
2883 'callback(1, -1, cb_arg)' is called when a number has passed the
2884 trial division stage.
2887 *) Fix for bug in CRL encoding. The validity dates weren't being handled
2891 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
2894 *) New function BN_pseudo_rand().
2897 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
2898 bignum version of BN_from_montgomery() with the working code from
2899 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
2903 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
2904 made it impossible to use the same SSL_SESSION data structure in
2905 SSL2 clients in multiple threads.
2908 *) The return value of RAND_load_file() no longer counts bytes obtained
2909 by stat(). RAND_load_file(..., -1) is new and uses the complete file
2910 to seed the PRNG (previously an explicit byte count was required).
2911 [Ulf Möller, Bodo Möller]
2913 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
2914 used (char *) instead of (void *) and had casts all over the place.
2917 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
2920 *) Retain source code compatibility for BN_prime_checks macro:
2921 BN_is_prime(..., BN_prime_checks, ...) now uses
2922 BN_prime_checks_for_size to determine the appropriate number of
2923 Rabin-Miller iterations.
2926 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
2927 DH_CHECK_P_NOT_SAFE_PRIME.
2928 (Check if this is true? OpenPGP calls them "strong".)
2931 *) Merge the functionality of "dh" and "gendh" programs into a new program
2932 "dhparam". The old programs are retained for now but will handle DH keys
2933 (instead of parameters) in future.
2936 *) Make the ciphers, s_server and s_client programs check the return values
2937 when a new cipher list is set.
2940 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
2941 ciphers. Before when the 56bit ciphers were enabled the sorting was
2944 The syntax for the cipher sorting has been extended to support sorting by
2945 cipher-strength (using the strength_bits hard coded in the tables).
2946 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
2948 Fix a bug in the cipher-command parser: when supplying a cipher command
2949 string with an "undefined" symbol (neither command nor alphanumeric
2950 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
2951 an error is flagged.
2953 Due to the strength-sorting extension, the code of the
2954 ssl_create_cipher_list() function was completely rearranged. I hope that
2955 the readability was also increased :-)
2956 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
2958 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
2959 for the first serial number and places 2 in the serial number file. This
2960 avoids problems when the root CA is created with serial number zero and
2961 the first user certificate has the same issuer name and serial number
2965 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
2966 the new code. Add documentation for this stuff.
2969 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
2970 X509_*() to X509at_*() on the grounds that they don't handle X509
2971 structures and behave in an analagous way to the X509v3 functions:
2972 they shouldn't be called directly but wrapper functions should be used
2975 So we also now have some wrapper functions that call the X509at functions
2976 when passed certificate requests. (TO DO: similar things can be done with
2977 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
2978 things. Some of these need some d2i or i2d and print functionality
2979 because they handle more complex structures.)
2982 *) Add missing #ifndefs that caused missing symbols when building libssl
2983 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
2984 NO_RSA in ssl/s2*.c.
2985 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
2987 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
2988 has a return value which indicates the quality of the random data
2989 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
2990 error queue. New function RAND_pseudo_bytes() generates output that is
2991 guaranteed to be unique but not unpredictable. RAND_add is like
2992 RAND_seed, but takes an extra argument for an entropy estimate
2993 (RAND_seed always assumes full entropy).
2996 *) Do more iterations of Rabin-Miller probable prime test (specifically,
2997 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
2998 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
2999 in crypto/bn/bn_prime.c for the complete table). This guarantees a
3000 false-positive rate of at most 2^-80 for random input.
3003 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
3006 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
3007 in the 0.9.5 release), this returns the chain
3008 from an X509_CTX structure with a dup of the stack and all
3009 the X509 reference counts upped: so the stack will exist
3010 after X509_CTX_cleanup() has been called. Modify pkcs12.c
3013 Also make SSL_SESSION_print() print out the verify return
3017 *) Add manpage for the pkcs12 command. Also change the default
3018 behaviour so MAC iteration counts are used unless the new
3019 -nomaciter option is used. This improves file security and
3020 only older versions of MSIE (4.0 for example) need it.
3023 *) Honor the no-xxx Configure options when creating .DEF files.
3026 *) Add PKCS#10 attributes to field table: challengePassword,
3027 unstructuredName and unstructuredAddress. These are taken from
3028 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
3029 international characters are used.
3031 More changes to X509_ATTRIBUTE code: allow the setting of types
3032 based on strings. Remove the 'loc' parameter when adding
3033 attributes because these will be a SET OF encoding which is sorted
3037 *) Initial changes to the 'req' utility to allow request generation
3038 automation. This will allow an application to just generate a template
3039 file containing all the field values and have req construct the
3042 Initial support for X509_ATTRIBUTE handling. Stacks of these are
3043 used all over the place including certificate requests and PKCS#7
3044 structures. They are currently handled manually where necessary with
3045 some primitive wrappers for PKCS#7. The new functions behave in a
3046 manner analogous to the X509 extension functions: they allow
3047 attributes to be looked up by NID and added.
3049 Later something similar to the X509V3 code would be desirable to
3050 automatically handle the encoding, decoding and printing of the
3051 more complex types. The string types like challengePassword can
3052 be handled by the string table functions.
3054 Also modified the multi byte string table handling. Now there is
3055 a 'global mask' which masks out certain types. The table itself
3056 can use the flag STABLE_NO_MASK to ignore the mask setting: this
3057 is useful when for example there is only one permissible type
3058 (as in countryName) and using the mask might result in no valid
3062 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
3063 SSL_get_peer_finished to allow applications to obtain the latest
3064 Finished messages sent to the peer or expected from the peer,
3065 respectively. (SSL_get_peer_finished is usually the Finished message
3066 actually received from the peer, otherwise the protocol will be aborted.)
3068 As the Finished message are message digests of the complete handshake
3069 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
3070 be used for external authentication procedures when the authentication
3071 provided by SSL/TLS is not desired or is not enough.
3074 *) Enhanced support for Alpha Linux is added. Now ./config checks if
3075 the host supports BWX extension and if Compaq C is present on the
3076 $PATH. Just exploiting of the BWX extension results in 20-30%
3077 performance kick for some algorithms, e.g. DES and RC4 to mention
3078 a couple. Compaq C in turn generates ~20% faster code for MD5 and
3082 *) Add support for MS "fast SGC". This is arguably a violation of the
3083 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
3084 weak crypto and after checking the certificate is SGC a second one
3085 with strong crypto. MS SGC stops the first handshake after receiving
3086 the server certificate message and sends a second client hello. Since
3087 a server will typically do all the time consuming operations before
3088 expecting any further messages from the client (server key exchange
3089 is the most expensive) there is little difference between the two.
3091 To get OpenSSL to support MS SGC we have to permit a second client
3092 hello message after we have sent server done. In addition we have to
3093 reset the MAC if we do get this second client hello.
3096 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
3097 if a DER encoded private key is RSA or DSA traditional format. Changed
3098 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
3099 format DER encoded private key. Newer code should use PKCS#8 format which
3100 has the key type encoded in the ASN1 structure. Added DER private key
3101 support to pkcs8 application.
3104 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
3105 ciphersuites has been selected (as required by the SSL 3/TLS 1
3106 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
3107 is set, we interpret this as a request to violate the specification
3108 (the worst that can happen is a handshake failure, and 'correct'
3109 behaviour would result in a handshake failure anyway).
3112 *) In SSL_CTX_add_session, take into account that there might be multiple
3113 SSL_SESSION structures with the same session ID (e.g. when two threads
3114 concurrently obtain them from an external cache).
3115 The internal cache can handle only one SSL_SESSION with a given ID,
3116 so if there's a conflict, we now throw out the old one to achieve
3120 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
3121 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
3122 some routines that use cipher OIDs: some ciphers do not have OIDs
3123 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
3127 *) Simplify the trust setting structure and code. Now we just have
3128 two sequences of OIDs for trusted and rejected settings. These will
3129 typically have values the same as the extended key usage extension
3130 and any application specific purposes.
3132 The trust checking code now has a default behaviour: it will just
3133 check for an object with the same NID as the passed id. Functions can
3134 be provided to override either the default behaviour or the behaviour
3135 for a given id. SSL client, server and email already have functions
3136 in place for compatibility: they check the NID and also return "trusted"
3137 if the certificate is self signed.
3140 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
3141 traditional format into an EVP_PKEY structure.
3144 *) Add a password callback function PEM_cb() which either prompts for
3145 a password if usr_data is NULL or otherwise assumes it is a null
3146 terminated password. Allow passwords to be passed on command line
3147 environment or config files in a few more utilities.
3150 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
3151 keys. Add some short names for PKCS#8 PBE algorithms and allow them
3152 to be specified on the command line for the pkcs8 and pkcs12 utilities.
3153 Update documentation.
3156 *) Support for ASN1 "NULL" type. This could be handled before by using
3157 ASN1_TYPE but there wasn't any function that would try to read a NULL
3158 and produce an error if it couldn't. For compatibility we also have
3159 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
3160 don't allocate anything because they don't need to.
3163 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
3165 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
3167 *) Rebuild of the memory allocation routines used by OpenSSL code and
3168 possibly others as well. The purpose is to make an interface that
3169 provide hooks so anyone can build a separate set of allocation and
3170 deallocation routines to be used by OpenSSL, for example memory
3171 pool implementations, or something else, which was previously hard
3172 since Malloc(), Realloc() and Free() were defined as macros having
3173 the values malloc, realloc and free, respectively (except for Win32
3174 compilations). The same is provided for memory debugging code.
3175 OpenSSL already comes with functionality to find memory leaks, but
3176 this gives people a chance to debug other memory problems.
3178 With these changes, a new set of functions and macros have appeared:
3180 CRYPTO_set_mem_debug_functions() [F]
3181 CRYPTO_get_mem_debug_functions() [F]
3182 CRYPTO_dbg_set_options() [F]
3183 CRYPTO_dbg_get_options() [F]
3184 CRYPTO_malloc_debug_init() [M]
3186 The memory debug functions are NULL by default, unless the library
3187 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
3188 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
3189 gives the standard debugging functions that come with OpenSSL) or
3190 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
3191 provided by the library user) must be used. When the standard
3192 debugging functions are used, CRYPTO_dbg_set_options can be used to
3193 request additional information:
3194 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
3195 the CRYPTO_MDEBUG_xxx macro when compiling the library.
3197 Also, things like CRYPTO_set_mem_functions will always give the
3198 expected result (the new set of functions is used for allocation
3199 and deallocation) at all times, regardless of platform and compiler
3202 To finish it up, some functions that were never use in any other
3203 way than through macros have a new API and new semantic:
3206 CRYPTO_dbg_realloc()
3209 All macros of value have retained their old syntax.
3210 [Richard Levitte and Bodo Moeller]
3212 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
3213 ordering of SMIMECapabilities wasn't in "strength order" and there
3214 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
3218 *) Some ASN1 types with illegal zero length encoding (INTEGER,
3219 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
3220 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
3222 *) Merge in my S/MIME library for OpenSSL. This provides a simple
3223 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
3224 functionality to handle multipart/signed properly) and a utility
3225 called 'smime' to call all this stuff. This is based on code I
3226 originally wrote for Celo who have kindly allowed it to be
3227 included in OpenSSL.
3230 *) Add variants des_set_key_checked and des_set_key_unchecked of
3231 des_set_key (aka des_key_sched). Global variable des_check_key
3232 decides which of these is called by des_set_key; this way
3233 des_check_key behaves as it always did, but applications and
3234 the library itself, which was buggy for des_check_key == 1,
3235 have a cleaner way to pick the version they need.
3238 *) New function PKCS12_newpass() which changes the password of a
3242 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
3243 dynamic mix. In both cases the ids can be used as an index into the
3244 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
3245 functions so they accept a list of the field values and the
3246 application doesn't need to directly manipulate the X509_TRUST
3250 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
3254 *) Modify the way the V3 extension code looks up extensions. This now
3255 works in a similar way to the object code: we have some "standard"
3256 extensions in a static table which is searched with OBJ_bsearch()
3257 and the application can add dynamic ones if needed. The file
3258 crypto/x509v3/ext_dat.h now has the info: this file needs to be
3259 updated whenever a new extension is added to the core code and kept
3260 in ext_nid order. There is a simple program 'tabtest.c' which checks
3261 this. New extensions are not added too often so this file can readily
3262 be maintained manually.
3264 There are two big advantages in doing things this way. The extensions
3265 can be looked up immediately and no longer need to be "added" using
3266 X509V3_add_standard_extensions(): this function now does nothing.
3267 [Side note: I get *lots* of email saying the extension code doesn't
3268 work because people forget to call this function]
3269 Also no dynamic allocation is done unless new extensions are added:
3270 so if we don't add custom extensions there is no need to call
3271 X509V3_EXT_cleanup().
3274 *) Modify enc utility's salting as follows: make salting the default. Add a
3275 magic header, so unsalted files fail gracefully instead of just decrypting
3276 to garbage. This is because not salting is a big security hole, so people
3277 should be discouraged from doing it.
3280 *) Fixes and enhancements to the 'x509' utility. It allowed a message
3281 digest to be passed on the command line but it only used this
3282 parameter when signing a certificate. Modified so all relevant
3283 operations are affected by the digest parameter including the
3284 -fingerprint and -x509toreq options. Also -x509toreq choked if a
3285 DSA key was used because it didn't fix the digest.
3288 *) Initial certificate chain verify code. Currently tests the untrusted
3289 certificates for consistency with the verify purpose (which is set
3290 when the X509_STORE_CTX structure is set up) and checks the pathlength.
3292 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
3293 this is because it will reject chains with invalid extensions whereas
3294 every previous version of OpenSSL and SSLeay made no checks at all.
3296 Trust code: checks the root CA for the relevant trust settings. Trust
3297 settings have an initial value consistent with the verify purpose: e.g.
3298 if the verify purpose is for SSL client use it expects the CA to be
3299 trusted for SSL client use. However the default value can be changed to
3300 permit custom trust settings: one example of this would be to only trust
3301 certificates from a specific "secure" set of CAs.
3303 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
3304 which should be used for version portability: especially since the
3305 verify structure is likely to change more often now.
3307 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
3308 to set them. If not set then assume SSL clients will verify SSL servers
3311 Two new options to the verify program: -untrusted allows a set of
3312 untrusted certificates to be passed in and -purpose which sets the
3313 intended purpose of the certificate. If a purpose is set then the
3314 new chain verify code is used to check extension consistency.
3317 *) Support for the authority information access extension.
3320 *) Modify RSA and DSA PEM read routines to transparently handle
3321 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
3322 public keys in a format compatible with certificate
3323 SubjectPublicKeyInfo structures. Unfortunately there were already
3324 functions called *_PublicKey_* which used various odd formats so
3325 these are retained for compatibility: however the DSA variants were
3326 never in a public release so they have been deleted. Changed dsa/rsa
3327 utilities to handle the new format: note no releases ever handled public
3328 keys so we should be OK.
3330 The primary motivation for this change is to avoid the same fiasco
3331 that dogs private keys: there are several incompatible private key
3332 formats some of which are standard and some OpenSSL specific and
3333 require various evil hacks to allow partial transparent handling and
3334 even then it doesn't work with DER formats. Given the option anything
3335 other than PKCS#8 should be dumped: but the other formats have to
3336 stay in the name of compatibility.
3338 With public keys and the benefit of hindsight one standard format
3339 is used which works with EVP_PKEY, RSA or DSA structures: though
3340 it clearly returns an error if you try to read the wrong kind of key.
3342 Added a -pubkey option to the 'x509' utility to output the public key.
3343 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
3344 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
3345 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
3346 that do the same as the EVP_PKEY_assign_*() except they up the
3347 reference count of the added key (they don't "swallow" the
3351 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
3352 CRLs would fail if the file contained no certificates or no CRLs:
3353 added a new function to read in both types and return the number
3354 read: this means that if none are read it will be an error. The
3355 DER versions of the certificate and CRL reader would always fail
3356 because it isn't possible to mix certificates and CRLs in DER format
3357 without choking one or the other routine. Changed this to just read
3358 a certificate: this is the best we can do. Also modified the code
3359 in apps/verify.c to take notice of return codes: it was previously
3360 attempting to read in certificates from NULL pointers and ignoring
3361 any errors: this is one reason why the cert and CRL reader seemed
3362 to work. It doesn't check return codes from the default certificate
3363 routines: these may well fail if the certificates aren't installed.
3366 *) Code to support otherName option in GeneralName.
3369 *) First update to verify code. Change the verify utility
3370 so it warns if it is passed a self signed certificate:
3371 for consistency with the normal behaviour. X509_verify
3372 has been modified to it will now verify a self signed
3373 certificate if *exactly* the same certificate appears
3374 in the store: it was previously impossible to trust a
3375 single self signed certificate. This means that:
3376 openssl verify ss.pem
3377 now gives a warning about a self signed certificate but
3378 openssl verify -CAfile ss.pem ss.pem
3382 *) For servers, store verify_result in SSL_SESSION data structure
3383 (and add it to external session representation).
3384 This is needed when client certificate verifications fails,
3385 but an application-provided verification callback (set by
3386 SSL_CTX_set_cert_verify_callback) allows accepting the session
3387 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
3388 but returns 1): When the session is reused, we have to set
3389 ssl->verify_result to the appropriate error code to avoid
3391 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
3393 *) Fix a bug in the new PKCS#7 code: it didn't consider the
3394 case in PKCS7_dataInit() where the signed PKCS7 structure
3395 didn't contain any existing data because it was being created.
3396 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
3398 *) Add a salt to the key derivation routines in enc.c. This
3399 forms the first 8 bytes of the encrypted file. Also add a
3400 -S option to allow a salt to be input on the command line.
3403 *) New function X509_cmp(). Oddly enough there wasn't a function
3404 to compare two certificates. We do this by working out the SHA1
3405 hash and comparing that. X509_cmp() will be needed by the trust
3409 *) SSL_get1_session() is like SSL_get_session(), but increments
3410 the reference count in the SSL_SESSION returned.
3411 [Geoff Thorpe <geoff@eu.c2.net>]
3413 *) Fix for 'req': it was adding a null to request attributes.
3414 Also change the X509_LOOKUP and X509_INFO code to handle
3415 certificate auxiliary information.
3418 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
3422 *) Add the possibility to add extra information to the memory leak
3423 detecting output, to form tracebacks, showing from where each
3424 allocation was originated: CRYPTO_push_info("constant string") adds
3425 the string plus current file name and line number to a per-thread
3426 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
3427 is like calling CYRPTO_pop_info() until the stack is empty.
3428 Also updated memory leak detection code to be multi-thread-safe.
3431 *) Add options -text and -noout to pkcs7 utility and delete the
3432 encryption options which never did anything. Update docs.
3435 *) Add options to some of the utilities to allow the pass phrase
3436 to be included on either the command line (not recommended on
3437 OSes like Unix) or read from the environment. Update the
3438 manpages and fix a few bugs.
3441 *) Add a few manpages for some of the openssl commands.
3444 *) Fix the -revoke option in ca. It was freeing up memory twice,
3445 leaking and not finding already revoked certificates.
3448 *) Extensive changes to support certificate auxiliary information.
3449 This involves the use of X509_CERT_AUX structure and X509_AUX
3450 functions. An X509_AUX function such as PEM_read_X509_AUX()
3451 can still read in a certificate file in the usual way but it
3452 will also read in any additional "auxiliary information". By
3453 doing things this way a fair degree of compatibility can be
3454 retained: existing certificates can have this information added
3455 using the new 'x509' options.
3457 Current auxiliary information includes an "alias" and some trust
3458 settings. The trust settings will ultimately be used in enhanced
3459 certificate chain verification routines: currently a certificate
3460 can only be trusted if it is self signed and then it is trusted
3464 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
3465 The problem was that one of the replacement routines had not been working
3466 since SSLeay releases. For now the offending routine has been replaced
3467 with non-optimised assembler. Even so, this now gives around 95%
3468 performance improvement for 1024 bit RSA signs.
3471 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
3472 handling. Most clients have the effective key size in bits equal to
3473 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
3474 A few however don't do this and instead use the size of the decrypted key
3475 to determine the RC2 key length and the AlgorithmIdentifier to determine
3476 the effective key length. In this case the effective key length can still
3477 be 40 bits but the key length can be 168 bits for example. This is fixed
3478 by manually forcing an RC2 key into the EVP_PKEY structure because the
3479 EVP code can't currently handle unusual RC2 key sizes: it always assumes
3480 the key length and effective key length are equal.
3483 *) Add a bunch of functions that should simplify the creation of
3484 X509_NAME structures. Now you should be able to do:
3485 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
3486 and have it automatically work out the correct field type and fill in
3487 the structures. The more adventurous can try:
3488 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
3489 and it will (hopefully) work out the correct multibyte encoding.
3492 *) Change the 'req' utility to use the new field handling and multibyte
3493 copy routines. Before the DN field creation was handled in an ad hoc
3494 way in req, ca, and x509 which was rather broken and didn't support
3495 BMPStrings or UTF8Strings. Since some software doesn't implement
3496 BMPStrings or UTF8Strings yet, they can be enabled using the config file
3497 using the dirstring_type option. See the new comment in the default
3498 openssl.cnf for more info.
3501 *) Make crypto/rand/md_rand.c more robust:
3502 - Assure unique random numbers after fork().
3503 - Make sure that concurrent threads access the global counter and
3504 md serializably so that we never lose entropy in them
3505 or use exactly the same state in multiple threads.
3506 Access to the large state is not always serializable because
3507 the additional locking could be a performance killer, and
3508 md should be large enough anyway.
3511 *) New file apps/app_rand.c with commonly needed functionality
3512 for handling the random seed file.
3514 Use the random seed file in some applications that previously did not:
3516 dsaparam -genkey (which also ignored its '-rand' option),
3519 x509 (when signing).
3520 Except on systems with /dev/urandom, it is crucial to have a random
3521 seed file at least for key creation, DSA signing, and for DH exchanges;
3522 for RSA signatures we could do without one.
3524 gendh and gendsa (unlike genrsa) used to read only the first byte
3525 of each file listed in the '-rand' option. The function as previously
3526 found in genrsa is now in app_rand.c and is used by all programs
3527 that support '-rand'.
3530 *) In RAND_write_file, use mode 0600 for creating files;
3531 don't just chmod when it may be too late.
3534 *) Report an error from X509_STORE_load_locations
3535 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
3538 *) New function ASN1_mbstring_copy() this copies a string in either
3539 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
3540 into an ASN1_STRING type. A mask of permissible types is passed
3541 and it chooses the "minimal" type to use or an error if not type
3545 *) Add function equivalents to the various macros in asn1.h. The old
3546 macros are retained with an M_ prefix. Code inside the library can
3547 use the M_ macros. External code (including the openssl utility)
3548 should *NOT* in order to be "shared library friendly".
3551 *) Add various functions that can check a certificate's extensions
3552 to see if it usable for various purposes such as SSL client,
3553 server or S/MIME and CAs of these types. This is currently
3554 VERY EXPERIMENTAL but will ultimately be used for certificate chain
3555 verification. Also added a -purpose flag to x509 utility to
3556 print out all the purposes.
3559 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
3563 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
3564 for, obtain and decode and extension and obtain its critical flag.
3565 This allows all the necessary extension code to be handled in a
3566 single function call.
3569 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
3570 platforms. See crypto/rc4/rc4_enc.c for further details.
3573 *) New -noout option to asn1parse. This causes no output to be produced
3574 its main use is when combined with -strparse and -out to extract data
3575 from a file (which may not be in ASN.1 format).
3578 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
3579 when producing the local key id.
3580 [Richard Levitte <levitte@stacken.kth.se>]
3582 *) New option -dhparam in s_server. This allows a DH parameter file to be
3583 stated explicitly. If it is not stated then it tries the first server
3584 certificate file. The previous behaviour hard coded the filename
3588 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
3589 a public key to be input or output. For example:
3590 openssl rsa -in key.pem -pubout -out pubkey.pem
3591 Also added necessary DSA public key functions to handle this.
3594 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
3595 in the message. This was handled by allowing
3596 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
3597 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
3599 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
3600 to the end of the strings whereas this didn't. This would cause problems
3601 if strings read with d2i_ASN1_bytes() were later modified.
3602 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
3604 *) Fix for base64 decode bug. When a base64 bio reads only one line of
3605 data and it contains EOF it will end up returning an error. This is
3606 caused by input 46 bytes long. The cause is due to the way base64
3607 BIOs find the start of base64 encoded data. They do this by trying a
3608 trial decode on each line until they find one that works. When they
3609 do a flag is set and it starts again knowing it can pass all the
3610 data directly through the decoder. Unfortunately it doesn't reset
3611 the context it uses. This means that if EOF is reached an attempt
3612 is made to pass two EOFs through the context and this causes the
3613 resulting error. This can also cause other problems as well. As is
3614 usual with these problems it takes *ages* to find and the fix is
3615 trivial: move one line.
3616 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
3618 *) Ugly workaround to get s_client and s_server working under Windows. The
3619 old code wouldn't work because it needed to select() on sockets and the
3620 tty (for keypresses and to see if data could be written). Win32 only
3621 supports select() on sockets so we select() with a 1s timeout on the
3622 sockets and then see if any characters are waiting to be read, if none
3623 are present then we retry, we also assume we can always write data to
3624 the tty. This isn't nice because the code then blocks until we've
3625 received a complete line of data and it is effectively polling the
3626 keyboard at 1s intervals: however it's quite a bit better than not
3627 working at all :-) A dedicated Windows application might handle this
3628 with an event loop for example.
3631 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
3632 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
3633 will be called when RSA_sign() and RSA_verify() are used. This is useful
3634 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
3635 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
3636 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
3637 This necessitated the support of an extra signature type NID_md5_sha1
3638 for SSL signatures and modifications to the SSL library to use it instead
3639 of calling RSA_public_decrypt() and RSA_private_encrypt().
3642 *) Add new -verify -CAfile and -CApath options to the crl program, these
3643 will lookup a CRL issuers certificate and verify the signature in a
3644 similar way to the verify program. Tidy up the crl program so it
3645 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
3646 less strict. It will now permit CRL extensions even if it is not
3647 a V2 CRL: this will allow it to tolerate some broken CRLs.
3650 *) Initialize all non-automatic variables each time one of the openssl
3651 sub-programs is started (this is necessary as they may be started
3652 multiple times from the "OpenSSL>" prompt).
3653 [Lennart Bang, Bodo Moeller]
3655 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
3656 removing all other RSA functionality (this is what NO_RSA does). This
3657 is so (for example) those in the US can disable those operations covered
3658 by the RSA patent while allowing storage and parsing of RSA keys and RSA
3662 *) Non-copying interface to BIO pairs.
3663 (still largely untested)
3666 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
3667 ASCII string. This was handled independently in various places before.
3670 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
3671 UTF8 strings a character at a time.
3674 *) Use client_version from client hello to select the protocol
3675 (s23_srvr.c) and for RSA client key exchange verification
3676 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
3679 *) Add various utility functions to handle SPKACs, these were previously
3680 handled by poking round in the structure internals. Added new function
3681 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
3682 print, verify and generate SPKACs. Based on an original idea from
3683 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
3686 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
3689 *) Allow the config file extension section to be overwritten on the
3690 command line. Based on an original idea from Massimiliano Pala
3691 <madwolf@comune.modena.it>. The new option is called -extensions
3692 and can be applied to ca, req and x509. Also -reqexts to override
3693 the request extensions in req and -crlexts to override the crl extensions
3697 *) Add new feature to the SPKAC handling in ca. Now you can include
3698 the same field multiple times by preceding it by "XXXX." for example:
3701 this is the same syntax as used in the req config file.
3704 *) Allow certificate extensions to be added to certificate requests. These
3705 are specified in a 'req_extensions' option of the req section of the
3706 config file. They can be printed out with the -text option to req but
3707 are otherwise ignored at present.
3710 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
3711 data read consists of only the final block it would not decrypted because
3712 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
3713 A misplaced 'break' also meant the decrypted final block might not be
3714 copied until the next read.
3717 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
3718 a few extra parameters to the DH structure: these will be useful if
3719 for example we want the value of 'q' or implement X9.42 DH.
3722 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
3723 provides hooks that allow the default DSA functions or functions on a
3724 "per key" basis to be replaced. This allows hardware acceleration and
3725 hardware key storage to be handled without major modification to the
3726 library. Also added low level modexp hooks and CRYPTO_EX structure and
3727 associated functions.
3730 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
3731 as "read only": it can't be written to and the buffer it points to will
3732 not be freed. Reading from a read only BIO is much more efficient than
3733 a normal memory BIO. This was added because there are several times when
3734 an area of memory needs to be read from a BIO. The previous method was
3735 to create a memory BIO and write the data to it, this results in two
3736 copies of the data and an O(n^2) reading algorithm. There is a new
3737 function BIO_new_mem_buf() which creates a read only memory BIO from
3738 an area of memory. Also modified the PKCS#7 routines to use read only
3742 *) Bugfix: ssl23_get_client_hello did not work properly when called in
3743 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
3744 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
3745 but a retry condition occured while trying to read the rest.
3748 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
3749 NID_pkcs7_encrypted by default: this was wrong since this should almost
3750 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
3751 the encrypted data type: this is a more sensible place to put it and it
3752 allows the PKCS#12 code to be tidied up that duplicated this
3756 *) Changed obj_dat.pl script so it takes its input and output files on
3757 the command line. This should avoid shell escape redirection problems
3761 *) Initial support for certificate extension requests, these are included
3762 in things like Xenroll certificate requests. Included functions to allow
3763 extensions to be obtained and added.
3766 *) -crlf option to s_client and s_server for sending newlines as
3767 CRLF (as required by many protocols).
3770 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
3772 *) Install libRSAglue.a when OpenSSL is built with RSAref.
3773 [Ralf S. Engelschall]
3775 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
3776 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
3778 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
3782 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
3783 DH parameters/keys (q is lost during that conversion, but the resulting
3784 DH parameters contain its length).
3786 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
3787 much faster than DH_generate_parameters (which creates parameters
3788 where p = 2*q + 1), and also the smaller q makes DH computations
3789 much more efficient (160-bit exponentiation instead of 1024-bit
3790 exponentiation); so this provides a convenient way to support DHE
3791 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
3792 utter importance to use
3793 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
3795 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
3796 when such DH parameters are used, because otherwise small subgroup
3797 attacks may become possible!
3800 *) Avoid memory leak in i2d_DHparams.
3803 *) Allow the -k option to be used more than once in the enc program:
3804 this allows the same encrypted message to be read by multiple recipients.
3807 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
3808 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
3809 it will always use the numerical form of the OID, even if it has a short
3813 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
3814 method only got called if p,q,dmp1,dmq1,iqmp components were present,
3815 otherwise bn_mod_exp was called. In the case of hardware keys for example
3816 no private key components need be present and it might store extra data
3817 in the RSA structure, which cannot be accessed from bn_mod_exp.
3818 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
3819 private key operations.
3822 *) Added support for SPARC Linux.
3825 *) pem_password_cb function type incompatibly changed from
3826 typedef int pem_password_cb(char *buf, int size, int rwflag);
3828 ....(char *buf, int size, int rwflag, void *userdata);
3829 so that applications can pass data to their callbacks:
3830 The PEM[_ASN1]_{read,write}... functions and macros now take an
3831 additional void * argument, which is just handed through whenever
3832 the password callback is called.
3833 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
3835 New function SSL_CTX_set_default_passwd_cb_userdata.
3837 Compatibility note: As many C implementations push function arguments
3838 onto the stack in reverse order, the new library version is likely to
3839 interoperate with programs that have been compiled with the old
3840 pem_password_cb definition (PEM_whatever takes some data that
3841 happens to be on the stack as its last argument, and the callback
3842 just ignores this garbage); but there is no guarantee whatsoever that
3845 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
3846 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
3847 problems not only on Windows, but also on some Unix platforms.
3848 To avoid problematic command lines, these definitions are now in an
3849 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
3850 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
3853 *) MIPS III/IV assembler module is reimplemented.
3856 *) More DES library cleanups: remove references to srand/rand and
3857 delete an unused file.
3860 *) Add support for the the free Netwide assembler (NASM) under Win32,
3861 since not many people have MASM (ml) and it can be hard to obtain.
3862 This is currently experimental but it seems to work OK and pass all
3863 the tests. Check out INSTALL.W32 for info.
3866 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
3867 without temporary keys kept an extra copy of the server key,
3868 and connections with temporary keys did not free everything in case
3872 *) New function RSA_check_key and new openssl rsa option -check
3873 for verifying the consistency of RSA keys.
3874 [Ulf Moeller, Bodo Moeller]
3876 *) Various changes to make Win32 compile work:
3877 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
3878 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
3879 comparison" warnings.
3880 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
3883 *) Add a debugging option to PKCS#5 v2 key generation function: when
3884 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
3885 derived keys are printed to stderr.
3888 *) Copy the flags in ASN1_STRING_dup().
3889 [Roman E. Pavlov <pre@mo.msk.ru>]
3891 *) The x509 application mishandled signing requests containing DSA
3892 keys when the signing key was also DSA and the parameters didn't match.
3894 It was supposed to omit the parameters when they matched the signing key:
3895 the verifying software was then supposed to automatically use the CA's
3896 parameters if they were absent from the end user certificate.
3898 Omitting parameters is no longer recommended. The test was also
3899 the wrong way round! This was probably due to unusual behaviour in
3900 EVP_cmp_parameters() which returns 1 if the parameters match.
3901 This meant that parameters were omitted when they *didn't* match and
3902 the certificate was useless. Certificates signed with 'ca' didn't have
3904 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
3906 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
3907 The interface is as follows:
3908 Applications can use
3909 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
3910 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
3911 "off" is now the default.
3912 The library internally uses
3913 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
3914 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
3915 to disable memory-checking temporarily.
3917 Some inconsistent states that previously were possible (and were
3918 even the default) are now avoided.
3920 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
3921 with each memory chunk allocated; this is occasionally more helpful
3922 than just having a counter.
3924 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
3926 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
3930 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
3931 which largely parallels "options", but is for changing API behaviour,
3932 whereas "options" are about protocol behaviour.
3933 Initial "mode" flags are:
3935 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
3936 a single record has been written.
3937 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
3938 retries use the same buffer location.
3939 (But all of the contents must be
3943 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
3946 *) Fix problems with no-hmac etc.
3947 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
3949 *) New functions RSA_get_default_method(), RSA_set_method() and
3950 RSA_get_method(). These allows replacement of RSA_METHODs without having
3951 to mess around with the internals of an RSA structure.
3954 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
3955 Also really enable memory leak checks in openssl.c and in some
3957 [Chad C. Mulligan, Bodo Moeller]
3959 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
3960 up the length of negative integers. This has now been simplified to just
3961 store the length when it is first determined and use it later, rather
3962 than trying to keep track of where data is copied and updating it to
3964 [Steve Henson, reported by Brien Wheeler
3965 <bwheeler@authentica-security.com>]
3967 *) Add a new function PKCS7_signatureVerify. This allows the verification
3968 of a PKCS#7 signature but with the signing certificate passed to the
3969 function itself. This contrasts with PKCS7_dataVerify which assumes the
3970 certificate is present in the PKCS#7 structure. This isn't always the
3971 case: certificates can be omitted from a PKCS#7 structure and be
3972 distributed by "out of band" means (such as a certificate database).
3975 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
3976 function prototypes in pem.h, also change util/mkdef.pl to add the
3977 necessary function names.
3980 *) mk1mf.pl (used by Windows builds) did not properly read the
3981 options set by Configure in the top level Makefile, and Configure
3982 was not even able to write more than one option correctly.
3983 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
3986 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
3987 file to be loaded from a BIO or FILE pointer. The BIO version will
3988 for example allow memory BIOs to contain config info.
3991 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
3992 Whoever hopes to achieve shared-library compatibility across versions
3993 must use this, not the compile-time macro.
3994 (Exercise 0.9.4: Which is the minimum library version required by
3996 Note: All this applies only to multi-threaded programs, others don't
4000 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
4001 through a BIO pair triggered the default case, i.e.
4002 SSLerr(...,SSL_R_UNKNOWN_STATE).
4005 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
4006 can use the SSL library even if none of the specific BIOs is
4010 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
4011 for the encoded length.
4012 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
4014 *) Add initial documentation of the X509V3 functions.
4017 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
4018 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
4019 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
4020 secure PKCS#8 private key format with a high iteration count.
4023 *) Fix determination of Perl interpreter: A perl or perl5
4024 _directory_ in $PATH was also accepted as the interpreter.
4025 [Ralf S. Engelschall]
4027 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
4028 wrong with it but it was very old and did things like calling
4029 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
4033 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
4034 to use the new extension code.
4037 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
4038 with macros. This should make it easier to change their form, add extra
4039 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
4043 *) Add to configuration table a new entry that can specify an alternative
4044 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
4045 according to Mark Crispin <MRC@Panda.COM>.
4049 *) DES CBC did not update the IV. Weird.
4052 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
4053 Changing the behaviour of the former might break existing programs --
4054 where IV updating is needed, des_ncbc_encrypt can be used.
4057 *) When bntest is run from "make test" it drives bc to check its
4058 calculations, as well as internally checking them. If an internal check
4059 fails, it needs to cause bc to give a non-zero result or make test carries
4060 on without noticing the failure. Fixed.
4063 *) DES library cleanups.
4066 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
4067 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
4068 ciphers. NOTE: although the key derivation function has been verified
4069 against some published test vectors it has not been extensively tested
4070 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
4074 *) Instead of "mkdir -p", which is not fully portable, use new
4075 Perl script "util/mkdir-p.pl".
4078 *) Rewrite the way password based encryption (PBE) is handled. It used to
4079 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
4080 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
4081 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
4082 the 'parameter' field of the AlgorithmIdentifier is passed to the
4083 underlying key generation function so it must do its own ASN1 parsing.
4084 This has also changed the EVP_PBE_CipherInit() function which now has a
4085 'parameter' argument instead of literal salt and iteration count values
4086 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
4089 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
4090 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
4091 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
4092 KEY" because this clashed with PKCS#8 unencrypted string. Since this
4093 value was just used as a "magic string" and not used directly its
4094 value doesn't matter.
4097 *) Introduce some semblance of const correctness to BN. Shame C doesn't
4101 *) "linux-sparc64" configuration (ultrapenguin).
4102 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
4103 "linux-sparc" configuration.
4104 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
4106 *) config now generates no-xxx options for missing ciphers.
4109 *) Support the EBCDIC character set (work in progress).
4110 File ebcdic.c not yet included because it has a different license.
4111 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
4113 *) Support BS2000/OSD-POSIX.
4114 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
4116 *) Make callbacks for key generation use void * instead of char *.
4119 *) Make S/MIME samples compile (not yet tested).
4122 *) Additional typesafe stacks.
4125 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
4129 Changes between 0.9.3 and 0.9.3a [29 May 1999]
4131 *) New configuration variant "sco5-gcc".
4133 *) Updated some demos.
4134 [Sean O Riordain, Wade Scholine]
4136 *) Add missing BIO_free at exit of pkcs12 application.
4139 *) Fix memory leak in conf.c.
4142 *) Updates for Win32 to assembler version of MD5.
4145 *) Set #! path to perl in apps/der_chop to where we found it
4146 instead of using a fixed path.
4149 *) SHA library changes for irix64-mips4-cc.
4152 *) Improvements for VMS support.
4156 Changes between 0.9.2b and 0.9.3 [24 May 1999]
4158 *) Bignum library bug fix. IRIX 6 passes "make test" now!
4159 This also avoids the problems with SC4.2 and unpatched SC5.
4160 [Andy Polyakov <appro@fy.chalmers.se>]
4162 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
4163 These are required because of the typesafe stack would otherwise break
4164 existing code. If old code used a structure member which used to be STACK
4165 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
4166 sk_num or sk_value it would produce an error because the num, data members
4167 are not present in STACK_OF. Now it just produces a warning. sk_set
4168 replaces the old method of assigning a value to sk_value
4169 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
4170 that does this will no longer work (and should use sk_set instead) but
4171 this could be regarded as a "questionable" behaviour anyway.
4174 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
4175 correctly handle encrypted S/MIME data.
4178 *) Change type of various DES function arguments from des_cblock
4179 (which means, in function argument declarations, pointer to char)
4180 to des_cblock * (meaning pointer to array with 8 char elements),
4181 which allows the compiler to do more typechecking; it was like
4182 that back in SSLeay, but with lots of ugly casts.
4184 Introduce new type const_des_cblock.
4187 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
4188 problems: find RecipientInfo structure that matches recipient certificate
4189 and initialise the ASN1 structures properly based on passed cipher.
4192 *) Belatedly make the BN tests actually check the results.
4195 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
4196 to and from BNs: it was completely broken. New compilation option
4197 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
4198 key elements as negative integers.
4201 *) Reorganize and speed up MD5.
4202 [Andy Polyakov <appro@fy.chalmers.se>]
4205 [Richard Levitte <richard@levitte.org>]
4207 *) New option -out to asn1parse to allow the parsed structure to be
4208 output to a file. This is most useful when combined with the -strparse
4209 option to examine the output of things like OCTET STRINGS.
4212 *) Make SSL library a little more fool-proof by not requiring any longer
4213 that SSL_set_{accept,connect}_state be called before
4214 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
4215 in many applications because usually everything *appeared* to work as
4216 intended anyway -- now it really works as intended).
4219 *) Move openssl.cnf out of lib/.
4222 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
4223 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
4224 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
4225 [Ralf S. Engelschall]
4227 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
4228 handle PKCS#7 enveloped data properly.
4229 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
4231 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
4232 copying pointers. The cert_st handling is changed by this in
4233 various ways (and thus what used to be known as ctx->default_cert
4234 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
4235 any longer when s->cert does not give us what we need).
4236 ssl_cert_instantiate becomes obsolete by this change.
4237 As soon as we've got the new code right (possibly it already is?),
4238 we have solved a couple of bugs of the earlier code where s->cert
4239 was used as if it could not have been shared with other SSL structures.
4241 Note that using the SSL API in certain dirty ways now will result
4242 in different behaviour than observed with earlier library versions:
4243 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
4244 does not influence s as it used to.
4246 In order to clean up things more thoroughly, inside SSL_SESSION
4247 we don't use CERT any longer, but a new structure SESS_CERT
4248 that holds per-session data (if available); currently, this is
4249 the peer's certificate chain and, for clients, the server's certificate
4250 and temporary key. CERT holds only those values that can have
4251 meaningful defaults in an SSL_CTX.
4254 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
4255 from the internal representation. Various PKCS#7 fixes: remove some
4256 evil casts and set the enc_dig_alg field properly based on the signing
4260 *) Allow PKCS#12 password to be set from the command line or the
4261 environment. Let 'ca' get its config file name from the environment
4262 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
4266 *) Allow certificate policies extension to use an IA5STRING for the
4267 organization field. This is contrary to the PKIX definition but
4268 VeriSign uses it and IE5 only recognises this form. Document 'x509'
4272 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
4273 without disallowing inline assembler and the like for non-pedantic builds.
4276 *) Support Borland C++ builder.
4277 [Janez Jere <jj@void.si>, modified by Ulf Möller]
4282 *) SHA-1 cleanups and performance enhancements.
4283 [Andy Polyakov <appro@fy.chalmers.se>]
4285 *) Sparc v8plus assembler for the bignum library.
4286 [Andy Polyakov <appro@fy.chalmers.se>]
4288 *) Accept any -xxx and +xxx compiler options in Configure.
4291 *) Update HPUX configuration.
4294 *) Add missing sk_<type>_unshift() function to safestack.h
4295 [Ralf S. Engelschall]
4297 *) New function SSL_CTX_use_certificate_chain_file that sets the
4298 "extra_cert"s in addition to the certificate. (This makes sense
4299 only for "PEM" format files, as chains as a whole are not
4303 *) Support verify_depth from the SSL API.
4304 x509_vfy.c had what can be considered an off-by-one-error:
4305 Its depth (which was not part of the external interface)
4306 was actually counting the number of certificates in a chain;
4307 now it really counts the depth.
4310 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
4311 instead of X509err, which often resulted in confusing error
4312 messages since the error codes are not globally unique
4313 (e.g. an alleged error in ssl3_accept when a certificate
4314 didn't match the private key).
4316 *) New function SSL_CTX_set_session_id_context that allows to set a default
4317 value (so that you don't need SSL_set_session_id_context for each
4318 connection using the SSL_CTX).
4321 *) OAEP decoding bug fix.
4324 *) Support INSTALL_PREFIX for package builders, as proposed by
4328 *) New Configure options "threads" and "no-threads". For systems
4329 where the proper compiler options are known (currently Solaris
4330 and Linux), "threads" is the default.
4333 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
4336 *) Install various scripts to $(OPENSSLDIR)/misc, not to
4337 $(INSTALLTOP)/bin -- they shouldn't clutter directories
4338 such as /usr/local/bin.
4341 *) "make linux-shared" to build shared libraries.
4342 [Niels Poppe <niels@netbox.org>]
4344 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
4347 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
4348 extension adding in x509 utility.
4351 *) Remove NOPROTO sections and error code comments.
4354 *) Partial rewrite of the DEF file generator to now parse the ANSI
4358 *) New Configure options --prefix=DIR and --openssldir=DIR.
4361 *) Complete rewrite of the error code script(s). It is all now handled
4362 by one script at the top level which handles error code gathering,
4363 header rewriting and C source file generation. It should be much better
4364 than the old method: it now uses a modified version of Ulf's parser to
4365 read the ANSI prototypes in all header files (thus the old K&R definitions
4366 aren't needed for error creation any more) and do a better job of
4367 translating function codes into names. The old 'ASN1 error code imbedded
4368 in a comment' is no longer necessary and it doesn't use .err files which
4369 have now been deleted. Also the error code call doesn't have to appear all
4370 on one line (which resulted in some large lines...).
4373 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
4376 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
4377 0 (which usually indicates a closed connection), but continue reading.
4380 *) Fix some race conditions.
4383 *) Add support for CRL distribution points extension. Add Certificate
4384 Policies and CRL distribution points documentation.
4387 *) Move the autogenerated header file parts to crypto/opensslconf.h.
4390 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
4391 8 of keying material. Merlin has also confirmed interop with this fix
4392 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
4393 [Merlin Hughes <merlin@baltimore.ie>]
4395 *) Fix lots of warnings.
4396 [Richard Levitte <levitte@stacken.kth.se>]
4398 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
4399 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
4400 [Richard Levitte <levitte@stacken.kth.se>]
4402 *) Fix problems with sizeof(long) == 8.
4403 [Andy Polyakov <appro@fy.chalmers.se>]
4405 *) Change functions to ANSI C.
4408 *) Fix typos in error codes.
4409 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
4411 *) Remove defunct assembler files from Configure.
4414 *) SPARC v8 assembler BIGNUM implementation.
4415 [Andy Polyakov <appro@fy.chalmers.se>]
4417 *) Support for Certificate Policies extension: both print and set.
4418 Various additions to support the r2i method this uses.
4421 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
4422 return a const string when you are expecting an allocated buffer.
4425 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
4426 types DirectoryString and DisplayText.
4429 *) Add code to allow r2i extensions to access the configuration database,
4430 add an LHASH database driver and add several ctx helper functions.
4433 *) Fix an evil bug in bn_expand2() which caused various BN functions to
4434 fail when they extended the size of a BIGNUM.
4437 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
4438 support typesafe stack.
4441 *) Fix typo in SSL_[gs]et_options().
4442 [Nils Frostberg <nils@medcom.se>]
4444 *) Delete various functions and files that belonged to the (now obsolete)
4445 old X509V3 handling code.
4448 *) New Configure option "rsaref".
4451 *) Don't auto-generate pem.h.
4454 *) Introduce type-safe ASN.1 SETs.
4457 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
4458 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
4460 *) Introduce type-safe STACKs. This will almost certainly break lots of code
4461 that links with OpenSSL (well at least cause lots of warnings), but fear
4462 not: the conversion is trivial, and it eliminates loads of evil casts. A
4463 few STACKed things have been converted already. Feel free to convert more.
4464 In the fullness of time, I'll do away with the STACK type altogether.
4467 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
4468 specified in <certfile> by updating the entry in the index.txt file.
4469 This way one no longer has to edit the index.txt file manually for
4470 revoking a certificate. The -revoke option does the gory details now.
4471 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
4473 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
4474 `-text' option at all and this way the `-noout -text' combination was
4475 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
4476 [Ralf S. Engelschall]
4478 *) Make sure a corresponding plain text error message exists for the
4479 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
4480 verify callback function determined that a certificate was revoked.
4481 [Ralf S. Engelschall]
4483 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
4484 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
4485 all available cipers including rc5, which was forgotten until now.
4486 In order to let the testing shell script know which algorithms
4487 are available, a new (up to now undocumented) command
4488 "openssl list-cipher-commands" is used.
4491 *) Bugfix: s_client occasionally would sleep in select() when
4492 it should have checked SSL_pending() first.
4495 *) New functions DSA_do_sign and DSA_do_verify to provide access to
4496 the raw DSA values prior to ASN.1 encoding.
4499 *) Tweaks to Configure
4500 [Niels Poppe <niels@netbox.org>]
4502 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
4506 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
4509 *) New config option to avoid instructions that are illegal on the 80386.
4510 The default code is faster, but requires at least a 486.
4513 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
4514 SSL2_SERVER_VERSION (not used at all) macros, which are now the
4515 same as SSL2_VERSION anyway.
4518 *) New "-showcerts" option for s_client.
4521 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
4522 application. Various cleanups and fixes.
4525 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
4526 modify error routines to work internally. Add error codes and PBE init
4527 to library startup routines.
4530 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
4531 packing functions to asn1 and evp. Changed function names and error
4532 codes along the way.
4535 *) PKCS12 integration: and so it begins... First of several patches to
4536 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
4537 objects to objects.h
4540 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
4541 and display support for Thawte strong extranet extension.
4544 *) Add LinuxPPC support.
4545 [Jeff Dubrule <igor@pobox.org>]
4547 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
4548 bn_div_words in alpha.s.
4549 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
4551 *) Make sure the RSA OAEP test is skipped under -DRSAref because
4552 OAEP isn't supported when OpenSSL is built with RSAref.
4553 [Ulf Moeller <ulf@fitug.de>]
4555 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
4556 so they no longer are missing under -DNOPROTO.
4557 [Soren S. Jorvang <soren@t.dk>]
4560 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
4562 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
4563 doesn't work when the session is reused. Coming soon!
4566 *) Fix a security hole, that allows sessions to be reused in the wrong
4567 context thus bypassing client cert protection! All software that uses
4568 client certs and session caches in multiple contexts NEEDS PATCHING to
4569 allow session reuse! A fuller solution is in the works.
4570 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
4572 *) Some more source tree cleanups (removed obsolete files
4573 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
4574 permission on "config" script to be executable) and a fix for the INSTALL
4576 [Ulf Moeller <ulf@fitug.de>]
4578 *) Remove some legacy and erroneous uses of malloc, free instead of
4580 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
4582 *) Make rsa_oaep_test return non-zero on error.
4583 [Ulf Moeller <ulf@fitug.de>]
4585 *) Add support for native Solaris shared libraries. Configure
4586 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
4587 if someone would make that last step automatic.
4588 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
4590 *) ctx_size was not built with the right compiler during "make links". Fixed.
4593 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
4594 except NULL ciphers". This means the default cipher list will no longer
4595 enable NULL ciphers. They need to be specifically enabled e.g. with
4596 the string "DEFAULT:eNULL".
4599 *) Fix to RSA private encryption routines: if p < q then it would
4600 occasionally produce an invalid result. This will only happen with
4601 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
4604 *) Be less restrictive and allow also `perl util/perlpath.pl
4605 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
4606 because this way one can also use an interpreter named `perl5' (which is
4607 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
4608 installed as `perl').
4609 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
4611 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
4612 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
4614 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
4615 advapi32.lib to Win32 build and change the pem test comparision
4616 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
4617 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
4618 and crypto/des/ede_cbcm_enc.c.
4621 *) DES quad checksum was broken on big-endian architectures. Fixed.
4624 *) Comment out two functions in bio.h that aren't implemented. Fix up the
4625 Win32 test batch file so it (might) work again. The Win32 test batch file
4626 is horrible: I feel ill....
4629 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
4630 in e_os.h. Audit of header files to check ANSI and non ANSI
4631 sections: 10 functions were absent from non ANSI section and not exported
4632 from Windows DLLs. Fixed up libeay.num for new functions.
4635 *) Make `openssl version' output lines consistent.
4636 [Ralf S. Engelschall]
4638 *) Fix Win32 symbol export lists for BIO functions: Added
4639 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
4640 to ms/libeay{16,32}.def.
4641 [Ralf S. Engelschall]
4643 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
4644 fine under Unix and passes some trivial tests I've now added. But the
4645 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
4646 added to make sure no one expects that this stuff really works in the
4647 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
4648 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
4650 [Ralf S. Engelschall]
4652 *) Fix the generation of two part addresses in perl.
4653 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
4655 *) Add config entry for Linux on MIPS.
4656 [John Tobey <jtobey@channel1.com>]
4658 *) Make links whenever Configure is run, unless we are on Windoze.
4661 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
4662 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
4666 *) Add a useful kludge to allow package maintainers to specify compiler and
4667 other platforms details on the command line without having to patch the
4668 Configure script everytime: One now can use ``perl Configure
4669 <id>:<details>'', i.e. platform ids are allowed to have details appended
4670 to them (seperated by colons). This is treated as there would be a static
4671 pre-configured entry in Configure's %table under key <id> with value
4672 <details> and ``perl Configure <id>'' is called. So, when you want to
4673 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
4674 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
4675 now, which overrides the FreeBSD-elf entry on-the-fly.
4676 [Ralf S. Engelschall]
4678 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
4681 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
4682 on the `perl Configure ...' command line. This way one can compile
4683 OpenSSL libraries with Position Independent Code (PIC) which is needed
4684 for linking it into DSOs.
4685 [Ralf S. Engelschall]
4687 *) Remarkably, export ciphers were totally broken and no-one had noticed!
4691 *) Cleaned up the LICENSE document: The official contact for any license
4692 questions now is the OpenSSL core team under openssl-core@openssl.org.
4693 And add a paragraph about the dual-license situation to make sure people
4694 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
4695 to the OpenSSL toolkit.
4696 [Ralf S. Engelschall]
4698 *) General source tree makefile cleanups: Made `making xxx in yyy...'
4699 display consistent in the source tree and replaced `/bin/rm' by `rm'.
4700 Additonally cleaned up the `make links' target: Remove unnecessary
4701 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
4702 to speed processing and no longer clutter the display with confusing
4703 stuff. Instead only the actually done links are displayed.
4704 [Ralf S. Engelschall]
4706 *) Permit null encryption ciphersuites, used for authentication only. It used
4707 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
4708 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
4712 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
4713 signed attributes when verifying signatures (this would break them),
4714 the detached data encoding was wrong and public keys obtained using
4715 X509_get_pubkey() weren't freed.
4718 *) Add text documentation for the BUFFER functions. Also added a work around
4719 to a Win95 console bug. This was triggered by the password read stuff: the
4720 last character typed gets carried over to the next fread(). If you were
4721 generating a new cert request using 'req' for example then the last
4722 character of the passphrase would be CR which would then enter the first
4726 *) Added the new `Includes OpenSSL Cryptography Software' button as
4727 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
4728 button and can be used by applications based on OpenSSL to show the
4729 relationship to the OpenSSL project.
4730 [Ralf S. Engelschall]
4732 *) Remove confusing variables in function signatures in files
4733 ssl/ssl_lib.c and ssl/ssl.h.
4734 [Lennart Bong <lob@kulthea.stacken.kth.se>]
4736 *) Don't install bss_file.c under PREFIX/include/
4737 [Lennart Bong <lob@kulthea.stacken.kth.se>]
4739 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
4740 functions that return function pointers and has support for NT specific
4741 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
4742 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
4743 unsigned to signed types: this was killing the Win32 compile.
4746 *) Add new certificate file to stack functions,
4747 SSL_add_dir_cert_subjects_to_stack() and
4748 SSL_add_file_cert_subjects_to_stack(). These largely supplant
4749 SSL_load_client_CA_file(), and can be used to add multiple certs easily
4750 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
4751 This means that Apache-SSL and similar packages don't have to mess around
4752 to add as many CAs as they want to the preferred list.
4755 *) Experiment with doxygen documentation. Currently only partially applied to
4757 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
4758 openssl.doxy as the configuration file.
4761 *) Get rid of remaining C++-style comments which strict C compilers hate.
4762 [Ralf S. Engelschall, pointed out by Carlos Amengual]
4764 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
4765 compiled in by default: it has problems with large keys.
4768 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
4769 DH private keys and/or callback functions which directly correspond to
4770 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
4771 is needed for applications which have to configure certificates on a
4772 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
4774 For the RSA certificate situation is makes no difference, but
4775 for the DSA certificate situation this fixes the "no shared cipher"
4776 problem where the OpenSSL cipher selection procedure failed because the
4777 temporary keys were not overtaken from the context and the API provided
4778 no way to reconfigure them.
4779 The new functions now let applications reconfigure the stuff and they
4780 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
4781 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
4782 non-public-API function ssl_cert_instantiate() is used as a helper
4783 function and also to reduce code redundancy inside ssl_rsa.c.
4784 [Ralf S. Engelschall]
4786 *) Move s_server -dcert and -dkey options out of the undocumented feature
4787 area because they are useful for the DSA situation and should be
4788 recognized by the users.
4789 [Ralf S. Engelschall]
4791 *) Fix the cipher decision scheme for export ciphers: the export bits are
4792 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
4793 SSL_EXP_MASK. So, the original variable has to be used instead of the
4794 already masked variable.
4795 [Richard Levitte <levitte@stacken.kth.se>]
4797 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
4798 [Richard Levitte <levitte@stacken.kth.se>]
4800 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
4801 from `int' to `unsigned int' because it's a length and initialized by
4802 EVP_DigestFinal() which expects an `unsigned int *'.
4803 [Richard Levitte <levitte@stacken.kth.se>]
4805 *) Don't hard-code path to Perl interpreter on shebang line of Configure
4806 script. Instead use the usual Shell->Perl transition trick.
4807 [Ralf S. Engelschall]
4809 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
4810 (in addition to RSA certificates) to match the behaviour of `openssl dsa
4811 -noout -modulus' as it's already the case for `openssl rsa -noout
4812 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
4813 currently the public key is printed (a decision which was already done by
4814 `openssl dsa -modulus' in the past) which serves a similar purpose.
4815 Additionally the NO_RSA no longer completely removes the whole -modulus
4816 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
4818 [Ralf S. Engelschall]
4820 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
4821 BIO. See the source (crypto/evp/bio_ok.c) for more info.
4822 [Arne Ansper <arne@ats.cyber.ee>]
4824 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
4825 to be added. Now both 'req' and 'ca' can use new objects defined in the
4829 *) Add cool BIO that does syslog (or event log on NT).
4830 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
4832 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
4833 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
4834 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
4835 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
4838 *) Add preliminary config info for new extension code.
4841 *) Make RSA_NO_PADDING really use no padding.
4842 [Ulf Moeller <ulf@fitug.de>]
4844 *) Generate errors when private/public key check is done.
4847 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
4848 for some CRL extensions and new objects added.
4851 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
4852 key usage extension and fuller support for authority key id.
4855 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
4856 padding method for RSA, which is recommended for new applications in PKCS
4857 #1 v2.0 (RFC 2437, October 1998).
4858 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
4859 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
4860 against Bleichbacher's attack on RSA.
4861 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
4864 *) Updates to the new SSL compression code
4865 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
4867 *) Fix so that the version number in the master secret, when passed
4868 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
4869 (because the server will not accept higher), that the version number
4870 is 0x03,0x01, not 0x03,0x00
4871 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
4873 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
4874 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
4875 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
4878 *) Support for RAW extensions where an arbitrary extension can be
4879 created by including its DER encoding. See apps/openssl.cnf for
4883 *) Make sure latest Perl versions don't interpret some generated C array
4884 code as Perl array code in the crypto/err/err_genc.pl script.
4885 [Lars Weber <3weber@informatik.uni-hamburg.de>]
4887 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
4888 not many people have the assembler. Various Win32 compilation fixes and
4889 update to the INSTALL.W32 file with (hopefully) more accurate Win32
4893 *) Modify configure script 'Configure' to automatically create crypto/date.h
4894 file under Win32 and also build pem.h from pem.org. New script
4895 util/mkfiles.pl to create the MINFO file on environments that can't do a
4896 'make files': perl util/mkfiles.pl >MINFO should work.
4899 *) Major rework of DES function declarations, in the pursuit of correctness
4900 and purity. As a result, many evil casts evaporated, and some weirdness,
4901 too. You may find this causes warnings in your code. Zapping your evil
4902 casts will probably fix them. Mostly.
4905 *) Fix for a typo in asn1.h. Bug fix to object creation script
4906 obj_dat.pl. It considered a zero in an object definition to mean
4907 "end of object": none of the objects in objects.h have any zeros
4908 so it wasn't spotted.
4909 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
4911 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
4912 Masking (CBCM). In the absence of test vectors, the best I have been able
4913 to do is check that the decrypt undoes the encrypt, so far. Send me test
4914 vectors if you have them.
4917 *) Correct calculation of key length for export ciphers (too much space was
4918 allocated for null ciphers). This has not been tested!
4921 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
4922 message is now correct (it understands "crypto" and "ssl" on its
4923 command line). There is also now an "update" option. This will update
4924 the util/ssleay.num and util/libeay.num files with any new functions.
4926 perl util/mkdef.pl crypto ssl update
4927 it will update them.
4930 *) Overhauled the Perl interface (perl/*):
4931 - ported BN stuff to OpenSSL's different BN library
4932 - made the perl/ source tree CVS-aware
4933 - renamed the package from SSLeay to OpenSSL (the files still contain
4934 their history because I've copied them in the repository)
4935 - removed obsolete files (the test scripts will be replaced
4936 by better Test::Harness variants in the future)
4937 [Ralf S. Engelschall]
4939 *) First cut for a very conservative source tree cleanup:
4940 1. merge various obsolete readme texts into doc/ssleay.txt
4941 where we collect the old documents and readme texts.
4942 2. remove the first part of files where I'm already sure that we no
4943 longer need them because of three reasons: either they are just temporary
4944 files which were left by Eric or they are preserved original files where
4945 I've verified that the diff is also available in the CVS via "cvs diff
4946 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
4947 the crypto/md/ stuff).
4948 [Ralf S. Engelschall]
4950 *) More extension code. Incomplete support for subject and issuer alt
4951 name, issuer and authority key id. Change the i2v function parameters
4952 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
4953 what that's for :-) Fix to ASN1 macro which messed up
4954 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
4957 *) Preliminary support for ENUMERATED type. This is largely copied from the
4961 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
4962 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
4964 *) Make sure `make rehash' target really finds the `openssl' program.
4965 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
4967 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
4968 like to hear about it if this slows down other processors.
4971 *) Add CygWin32 platform information to Configure script.
4972 [Alan Batie <batie@aahz.jf.intel.com>]
4974 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
4975 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
4977 *) New program nseq to manipulate netscape certificate sequences
4980 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
4984 *) Fixes to BN code. Previously the default was to define BN_RECURSION
4985 but the BN code had some problems that would cause failures when
4986 doing certificate verification and some other functions.
4987 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
4989 *) Add ASN1 and PEM code to support netscape certificate sequences.
4992 *) Add ASN1 and PEM code to support netscape certificate sequences.
4995 *) Add several PKIX and private extended key usage OIDs.
4998 *) Modify the 'ca' program to handle the new extension code. Modify
4999 openssl.cnf for new extension format, add comments.
5002 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
5003 and add a sample to openssl.cnf so req -x509 now adds appropriate
5007 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
5008 error code, add initial support to X509_print() and x509 application.
5011 *) Takes a deep breath and start addding X509 V3 extension support code. Add
5012 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
5013 stuff is currently isolated and isn't even compiled yet.
5016 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
5017 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
5018 Removed the versions check from X509 routines when loading extensions:
5019 this allows certain broken certificates that don't set the version
5020 properly to be processed.
5023 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
5024 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
5025 can still be regenerated with "make depend".
5028 *) Spelling mistake in C version of CAST-128.
5029 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
5031 *) Changes to the error generation code. The perl script err-code.pl
5032 now reads in the old error codes and retains the old numbers, only
5033 adding new ones if necessary. It also only changes the .err files if new
5034 codes are added. The makefiles have been modified to only insert errors
5035 when needed (to avoid needlessly modifying header files). This is done
5036 by only inserting errors if the .err file is newer than the auto generated
5037 C file. To rebuild all the error codes from scratch (the old behaviour)
5038 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
5039 or delete all the .err files.
5042 *) CAST-128 was incorrectly implemented for short keys. The C version has
5043 been fixed, but is untested. The assembler versions are also fixed, but
5044 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
5045 to regenerate it if needed.
5046 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
5047 Hagino <itojun@kame.net>]
5049 *) File was opened incorrectly in randfile.c.
5050 [Ulf Möller <ulf@fitug.de>]
5052 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
5053 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
5054 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
5055 al: it's just almost always a UTCTime. Note this patch adds new error
5056 codes so do a "make errors" if there are problems.
5059 *) Correct Linux 1 recognition in config.
5060 [Ulf Möller <ulf@fitug.de>]
5062 *) Remove pointless MD5 hash when using DSA keys in ca.
5063 [Anonymous <nobody@replay.com>]
5065 *) Generate an error if given an empty string as a cert directory. Also
5066 generate an error if handed NULL (previously returned 0 to indicate an
5067 error, but didn't set one).
5068 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
5070 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
5073 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
5074 parameters. This was causing a warning which killed off the Win32 compile.
5077 *) Remove C++ style comments from crypto/bn/bn_local.h.
5078 [Neil Costigan <neil.costigan@celocom.com>]
5080 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
5081 based on a text string, looking up short and long names and finally
5082 "dot" format. The "dot" format stuff didn't work. Added new function
5083 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
5084 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
5085 OID is not part of the table.
5088 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
5089 X509_LOOKUP_by_alias().
5092 *) Sort openssl functions by name.
5095 *) Get the gendsa program working (hopefully) and add it to app list. Remove
5096 encryption from sample DSA keys (in case anyone is interested the password
5100 *) Make _all_ *_free functions accept a NULL pointer.
5101 [Frans Heymans <fheymans@isaserver.be>]
5103 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
5105 [Anonymous <nobody@replay.com>]
5107 *) s_server should send the CAfile as acceptable CAs, not its own cert.
5108 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
5110 *) Don't blow it for numeric -newkey arguments to apps/req.
5111 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
5113 *) Temp key "for export" tests were wrong in s3_srvr.c.
5114 [Anonymous <nobody@replay.com>]
5116 *) Add prototype for temp key callback functions
5117 SSL_CTX_set_tmp_{rsa,dh}_callback().
5120 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
5121 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
5124 *) X509_name_add_entry() freed the wrong thing after an error.
5125 [Arne Ansper <arne@ats.cyber.ee>]
5127 *) rsa_eay.c would attempt to free a NULL context.
5128 [Arne Ansper <arne@ats.cyber.ee>]
5130 *) BIO_s_socket() had a broken should_retry() on Windoze.
5131 [Arne Ansper <arne@ats.cyber.ee>]
5133 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
5134 [Arne Ansper <arne@ats.cyber.ee>]
5136 *) Make sure the already existing X509_STORE->depth variable is initialized
5137 in X509_STORE_new(), but document the fact that this variable is still
5138 unused in the certificate verification process.
5139 [Ralf S. Engelschall]
5141 *) Fix the various library and apps files to free up pkeys obtained from
5142 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
5145 *) Fix reference counting in X509_PUBKEY_get(). This makes
5146 demos/maurice/example2.c work, amongst others, probably.
5147 [Steve Henson and Ben Laurie]
5149 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
5150 `openssl' and second, the shortcut symlinks for the `openssl <command>'
5151 are no longer created. This way we have a single and consistent command
5152 line interface `openssl <command>', similar to `cvs <command>'.
5153 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
5155 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
5156 BIT STRING wrapper always have zero unused bits.
5159 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
5162 *) Make the top-level INSTALL documentation easier to understand.
5165 *) Makefiles updated to exit if an error occurs in a sub-directory
5166 make (including if user presses ^C) [Paul Sutton]
5168 *) Make Montgomery context stuff explicit in RSA data structure.
5171 *) Fix build order of pem and err to allow for generated pem.h.
5174 *) Fix renumbering bug in X509_NAME_delete_entry().
5177 *) Enhanced the err-ins.pl script so it makes the error library number
5178 global and can add a library name. This is needed for external ASN1 and
5179 other error libraries.
5182 *) Fixed sk_insert which never worked properly.
5185 *) Fix ASN1 macros so they can handle indefinite length construted
5186 EXPLICIT tags. Some non standard certificates use these: they can now
5190 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
5191 into a single doc/ssleay.txt bundle. This way the information is still
5192 preserved but no longer messes up this directory. Now it's new room for
5193 the new set of documenation files.
5194 [Ralf S. Engelschall]
5196 *) SETs were incorrectly DER encoded. This was a major pain, because they
5197 shared code with SEQUENCEs, which aren't coded the same. This means that
5198 almost everything to do with SETs or SEQUENCEs has either changed name or
5199 number of arguments.
5200 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
5202 *) Fix test data to work with the above.
5205 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
5206 was already fixed by Eric for 0.9.1 it seems.
5207 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
5209 *) Autodetect FreeBSD3.
5212 *) Fix various bugs in Configure. This affects the following platforms:
5216 unixware-2.0-pentium
5220 *) Eliminate generated files from CVS. Reorder tests to regenerate files
5221 before they are needed.
5224 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
5228 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
5230 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
5231 changed SSLeay to OpenSSL in version strings.
5232 [Ralf S. Engelschall]
5234 *) Some fixups to the top-level documents.
5237 *) Fixed the nasty bug where rsaref.h was not found under compile-time
5238 because the symlink to include/ was missing.
5239 [Ralf S. Engelschall]
5241 *) Incorporated the popular no-RSA/DSA-only patches
5242 which allow to compile a RSA-free SSLeay.
5243 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
5245 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
5246 when "ssleay" is still not found.
5247 [Ralf S. Engelschall]
5249 *) Added more platforms to Configure: Cray T3E, HPUX 11,
5250 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
5252 *) Updated the README file.
5253 [Ralf S. Engelschall]
5255 *) Added various .cvsignore files in the CVS repository subdirs
5256 to make a "cvs update" really silent.
5257 [Ralf S. Engelschall]
5259 *) Recompiled the error-definition header files and added
5260 missing symbols to the Win32 linker tables.
5261 [Ralf S. Engelschall]
5263 *) Cleaned up the top-level documents;
5264 o new files: CHANGES and LICENSE
5265 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
5266 o merged COPYRIGHT into LICENSE
5267 o removed obsolete TODO file
5268 o renamed MICROSOFT to INSTALL.W32
5269 [Ralf S. Engelschall]
5271 *) Removed dummy files from the 0.9.1b source tree:
5272 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
5273 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
5274 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
5275 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
5276 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
5277 [Ralf S. Engelschall]
5279 *) Added various platform portability fixes.
5282 *) The Genesis of the OpenSSL rpject:
5283 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
5284 Young and Tim J. Hudson created while they were working for C2Net until
5286 [The OpenSSL Project]
5289 Changes between 0.9.0b and 0.9.1b [not released]
5291 *) Updated a few CA certificates under certs/
5294 *) Changed some BIGNUM api stuff.
5297 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
5298 DGUX x86, Linux Alpha, etc.
5301 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
5302 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
5306 *) Add -strparse option to asn1pars program which parses nested
5308 [Dr Stephen Henson <shenson@bigfoot.com>]
5310 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
5313 *) DSA fix for "ca" program.
5316 *) Added "-genkey" option to "dsaparam" program.
5319 *) Added RIPE MD160 (rmd160) message digest.
5322 *) Added -a (all) option to "ssleay version" command.
5325 *) Added PLATFORM define which is the id given to Configure.
5328 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
5331 *) Extended the ASN.1 parser routines.
5334 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
5337 *) Added a BN_CTX to the BN library.
5340 *) Fixed the weak key values in DES library
5343 *) Changed API in EVP library for cipher aliases.
5346 *) Added support for RC2/64bit cipher.
5349 *) Converted the lhash library to the crypto/mem.c functions.
5352 *) Added more recognized ASN.1 object ids.
5355 *) Added more RSA padding checks for SSL/TLS.
5358 *) Added BIO proxy/filter functionality.
5361 *) Added extra_certs to SSL_CTX which can be used
5362 send extra CA certificates to the client in the CA cert chain sending
5363 process. It can be configured with SSL_CTX_add_extra_chain_cert().
5366 *) Now Fortezza is denied in the authentication phase because
5367 this is key exchange mechanism is not supported by SSLeay at all.
5370 *) Additional PKCS1 checks.
5373 *) Support the string "TLSv1" for all TLS v1 ciphers.
5376 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
5377 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
5380 *) Fixed a few memory leaks.
5383 *) Fixed various code and comment typos.
5386 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
5387 bytes sent in the client random.
5388 [Edward Bishop <ebishop@spyglass.com>]
projects / openssl.git / blob