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 +) New function SSL_renegotiate_pending(). This returns true once
16 renegotiation has been requested (either SSL_renegotiate() call
17 or HelloRequest/ClientHello receveived from the peer) and becomes
18 false once a handshake has been completed.
19 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
20 sends a HelloRequest, but does not ensure that a handshake takes
21 place. SSL_renegotiate_pending() is useful for checking if the
22 client has followed the request.)
25 +) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
26 By default, clients may request session resumption even during
27 renegotiation (if session ID contexts permit); with this option,
28 session resumption is possible only in the first handshake.
31 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
32 should end in 'break', not 'goto end' which circuments various
33 cleanups done in state SSL_ST_OK. But session related stuff
34 must be disabled for SSL_ST_OK in the case that we just sent a
37 Also avoid some overhead by not calling ssl_init_wbio_buffer()
38 before just sending a HelloRequest.
39 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
41 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
42 reveal whether illegal block cipher padding was found or a MAC
43 verification error occured. (Neither SSLerr() codes nor alerts
44 are directly visible to potential attackers, but the information
45 may leak via logfiles.)
47 Similar changes are not required for the SSL 2.0 implementation
48 because the number of padding bytes is sent in clear for SSL 2.0,
49 and the extra bytes are just ignored. However ssl/s2_pkt.c
50 failed to verify that the purported number of padding bytes is in
54 +) Add some demos for certificate and certificate request creation.
57 +) Make maximum certificate chain size accepted from the peer application
58 settable (SSL*_get/set_max_cert_list()), as proposed by
59 "Douglas E. Engert" <deengert@anl.gov>.
62 +) Add support for shared libraries for Unixware-7 and support including
63 shared libraries for OpenUNIX-8 (Boyd Lynn Gerber <gerberb@zenez.com>).
66 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
67 'wristwatch attack' using huge encoding parameters (cf.
68 James H. Manger's CRYPTO 2001 paper). Note that the
69 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
70 encoding parameters and hence was not vulnerable.
73 +) Add a "destroy" handler to ENGINEs that allows structural cleanup to
74 be done prior to destruction. Use this to unload error strings from
75 ENGINEs that load their own error strings. NB: This adds two new API
76 functions to "get" and "set" this destroy handler in an ENGINE.
79 +) Alter all existing ENGINE implementations (except "openssl" and
80 "openbsd") to dynamically instantiate their own error strings. This
81 makes them more flexible to be built both as statically-linked ENGINEs
82 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
83 Also, add stub code to each that makes building them as self-contained
84 shared-libraries easier (see README.ENGINE).
87 +) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
88 implementations into applications that are completely implemented in
89 self-contained shared-libraries. The "dynamic" ENGINE exposes control
90 commands that can be used to configure what shared-library to load and
91 to control aspects of the way it is handled. Also, made an update to
92 the README.ENGINE file that brings its information up-to-date and
93 provides some information and instructions on the "dynamic" ENGINE
94 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
98 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
100 *) Make it possible to unload ranges of ERR strings with a new
101 "ERR_unload_strings" function.
104 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
105 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
106 followed by modular reduction.
107 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
109 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
110 equivalent based on BN_pseudo_rand() instead of BN_rand().
113 +) Add a copy() function to EVP_MD.
116 +) Make EVP_MD routines take a context pointer instead of just the
117 md_data void pointer.
120 +) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
121 that the digest can only process a single chunk of data
122 (typically because it is provided by a piece of
123 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
124 is only going to provide a single chunk of data, and hence the
125 framework needn't accumulate the data for oneshot drivers.
128 +) As with "ERR", make it possible to replace the underlying "ex_data"
129 functions. This change also alters the storage and management of global
130 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
131 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
132 index counters. The API functions that use this state have been changed
133 to take a "class_index" rather than pointers to the class's local STACK
134 and counter, and there is now an API function to dynamically create new
135 classes. This centralisation allows us to (a) plug a lot of the
136 thread-safety problems that existed, and (b) makes it possible to clean
137 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
138 such data would previously have always leaked in application code and
139 workarounds were in place to make the memory debugging turn a blind eye
140 to it. Application code that doesn't use this new function will still
141 leak as before, but their memory debugging output will announce it now
142 rather than letting it slide.
144 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
145 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
146 has a return value to indicate success or failure.
149 +) Make it possible to replace the underlying "ERR" functions such that the
150 global state (2 LHASH tables and 2 locks) is only used by the "default"
151 implementation. This change also adds two functions to "get" and "set"
152 the implementation prior to it being automatically set the first time
153 any other ERR function takes place. Ie. an application can call "get",
154 pass the return value to a module it has just loaded, and that module
155 can call its own "set" function using that value. This means the
156 module's "ERR" operations will use (and modify) the error state in the
157 application and not in its own statically linked copy of OpenSSL code.
160 +) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
161 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
162 the operation, and provides a more encapsulated way for external code
163 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
164 to use these functions rather than manually incrementing the counts.
166 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
169 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
170 This function was broken, as the check for a new client hello message
171 to handle SGC did not allow these large messages.
172 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
175 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
178 +) Add EVP test program.
181 +) Add symmetric cipher support to ENGINE. Expect the API to change!
184 +) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
185 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
186 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
187 These allow a CRL to be built without having to access X509_CRL fields
188 directly. Modify 'ca' application to use new functions.
191 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
192 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
195 *) Rework the configuration and shared library support for Tru64 Unix.
196 The configuration part makes use of modern compiler features and
197 still retains old compiler behavior for those that run older versions
198 of the OS. The shared library support part includes a variant that
199 uses the RPATH feature, and is available through the special
200 configuration target "alpha-cc-rpath", which will never be selected
202 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
204 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
205 with the same message size as in ssl3_get_certificate_request().
206 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
207 messages might inadvertently be reject as too long.
208 [Petr Lampa <lampa@fee.vutbr.cz>]
210 +) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
211 bug workarounds. Rollback attack detection is a security feature.
212 The problem will only arise on OpenSSL servers when TLSv1 is not
213 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
214 Software authors not wanting to support TLSv1 will have special reasons
215 for their choice and can explicitly enable this option.
216 [Bodo Moeller, Lutz Jaenicke]
218 +) Rationalise EVP so it can be extended: don't include a union of
219 cipher/digest structures, add init/cleanup functions. This also reduces
220 the number of header dependencies.
223 +) Make DES key schedule conform to the usual scheme, as well as
224 correcting its structure. This means that calls to DES functions
225 now have to pass a pointer to a des_key_schedule instead of a
226 plain des_key_schedule (which was actually always a pointer
230 +) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
233 *) Modified SSL library such that the verify_callback that has been set
234 specificly for an SSL object with SSL_set_verify() is actually being
235 used. Before the change, a verify_callback set with this function was
236 ignored and the verify_callback() set in the SSL_CTX at the time of
237 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
238 to allow the necessary settings.
241 +) Initial reduction of linker bloat: the use of some functions, such as
242 PEM causes large amounts of unused functions to be linked in due to
243 poor organisation. For example pem_all.c contains every PEM function
244 which has a knock on effect of linking in large amounts of (unused)
245 ASN1 code. Grouping together similar functions and splitting unrelated
246 functions prevents this.
249 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
250 explicitly to NULL, as at least on Solaris 8 this seems not always to be
251 done automatically (in contradiction to the requirements of the C
252 standard). This made problems when used from OpenSSH.
255 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
256 dh->length and always used
258 BN_rand_range(priv_key, dh->p).
260 BN_rand_range() is not necessary for Diffie-Hellman, and this
261 specific range makes Diffie-Hellman unnecessarily inefficient if
262 dh->length (recommended exponent length) is much smaller than the
263 length of dh->p. We could use BN_rand_range() if the order of
264 the subgroup was stored in the DH structure, but we only have
269 BN_rand(priv_key, l, ...)
271 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
277 RSA_eay_public_encrypt
278 RSA_eay_private_decrypt
279 RSA_eay_private_encrypt (signing)
280 RSA_eay_public_decrypt (signature verification)
282 (default implementations for RSA_public_encrypt,
283 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
284 always reject numbers >= n.
287 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
288 to synchronize access to 'locking_thread'. This is necessary on
289 systems where access to 'locking_thread' (an 'unsigned long'
290 variable) is not atomic.
293 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
294 *before* setting the 'crypto_lock_rand' flag. The previous code had
295 a race condition if 0 is a valid thread ID.
296 [Travis Vitek <vitek@roguewave.com>]
298 +) Cleanup of EVP macros.
301 +) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
305 +) Add initial OCSP responder support to ocsp application. The
306 revocation information is handled using the text based index
307 use by the ca application. The responder can either handle
308 requests generated internally, supplied in files (for example
309 via a CGI script) or using an internal minimal server.
312 +) Add configuration choices to get zlib compression for TLS.
315 +) Changes to Kerberos SSL for RFC 2712 compliance:
316 1. Implemented real KerberosWrapper, instead of just using
317 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
318 2. Implemented optional authenticator field of KerberosWrapper.
320 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
321 and authenticator structs; see crypto/krb5/.
323 Generalized Kerberos calls to support multiple Kerberos libraries.
324 [Vern Staats <staatsvr@asc.hpc.mil>,
325 Jeffrey Altman <jaltman@columbia.edu>
328 +) Cause 'openssl speed' to use fully hard-coded DSA keys as it
329 already does with RSA. testdsa.h now has 'priv_key/pub_key'
330 values for each of the key sizes rather than having just
331 parameters (and 'speed' generating keys each time).
334 -) OpenSSL 0.9.6b released [9 July 2001]
336 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
337 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
338 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
339 PRNG state recovery was possible based on the output of
340 one PRNG request appropriately sized to gain knowledge on
341 'md' followed by enough consecutive 1-byte PRNG requests
342 to traverse all of 'state'.
344 1. When updating 'md_local' (the current thread's copy of 'md')
345 during PRNG output generation, hash all of the previous
346 'md_local' value, not just the half used for PRNG output.
348 2. Make the number of bytes from 'state' included into the hash
349 independent from the number of PRNG bytes requested.
351 The first measure alone would be sufficient to avoid
352 Markku-Juhani's attack. (Actually it had never occurred
353 to me that the half of 'md_local' used for chaining was the
354 half from which PRNG output bytes were taken -- I had always
355 assumed that the secret half would be used.) The second
356 measure makes sure that additional data from 'state' is never
357 mixed into 'md_local' in small portions; this heuristically
358 further strengthens the PRNG.
361 +) Speed up EVP routines.
364 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
365 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
366 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
367 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
369 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
370 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
371 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
374 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
376 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
379 *) Fix crypto/bn/asm/mips3.s.
382 *) When only the key is given to "enc", the IV is undefined. Print out
383 an error message in this case.
386 +) Added the OS2-EMX target.
387 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
389 +) Rewrite apps to use NCONF routines instead of the old CONF. New functions
390 to support NCONF routines in extension code. New function CONF_set_nconf()
391 to allow functions which take an NCONF to also handle the old LHASH
392 structure: this means that the old CONF compatible routines can be
393 retained (in particular wrt extensions) without having to duplicate the
394 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
397 *) Handle special case when X509_NAME is empty in X509 printing routines.
400 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
401 positive and less than q.
404 +) Enhance the general user interface with mechanisms for inner control
405 and with possibilities to have yes/no kind of prompts.
408 +) Change all calls to low level digest routines in the library and
409 applications to use EVP. Add missing calls to HMAC_cleanup() and
410 don't assume HMAC_CTX can be copied using memcpy().
411 [Verdon Walker <VWalker@novell.com>, Steve Henson]
413 +) Add the possibility to control engines through control names but with
414 arbitrary arguments instead of just a string.
415 Change the key loaders to take a UI_METHOD instead of a callback
416 function pointer. NOTE: this breaks binary compatibility with earlier
417 versions of OpenSSL [engine].
418 Adapt the nCipher code for these new conditions and add a card insertion
422 +) Enhance the general user interface with mechanisms to better support
423 dialog box interfaces, application-defined prompts, the possibility
424 to use defaults (for example default passwords from somewhere else)
425 and interrupts/cancellations.
428 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
429 used: it isn't thread safe and the add_lock_callback should handle
431 [Paul Rose <Paul.Rose@bridge.com>]
433 *) Verify that incoming data obeys the block size in
434 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
437 +) Tidy up PKCS#12 attribute handling. Add support for the CSP name
438 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
442 [Ulf Möller, Bodo Möller]
444 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
445 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
446 when fixing the server behaviour for backwards-compatible 'client
447 hello' messages. (Note that the attack is impractical against
448 SSL 3.0 and TLS 1.0 anyway because length and version checking
449 means that the probability of guessing a valid ciphertext is
450 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
453 Before 0.9.5, the countermeasure (hide the error by generating a
454 random 'decryption result') did not work properly because
455 ERR_clear_error() was missing, meaning that SSL_get_error() would
456 detect the supposedly ignored error.
458 Both problems are now fixed.
461 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
462 (previously it was 1024).
465 +) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
466 tidy up some unnecessarily weird code in 'sk_new()').
467 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
469 +) Change the key loading routines for ENGINEs to use the same kind
470 callback (pem_password_cb) as all other routines that need this
474 *) Fix for compatibility mode trust settings: ignore trust settings
475 unless some valid trust or reject settings are present.
478 *) Fix for blowfish EVP: its a variable length cipher.
481 +) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
482 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
483 than this minimum value is recommended.
486 +) New random seeder for OpenVMS, using the system process statistics
487 that are easily reachable.
490 +) Windows apparently can't transparently handle global
491 variables defined in DLLs. Initialisations such as:
493 const ASN1_ITEM *it = &ASN1_INTEGER_it;
495 wont compile. This is used by the any applications that need to
496 declare their own ASN1 modules. This was fixed by adding the option
497 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
498 needed for static libraries under Win32.
501 +) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
502 setting of purpose and trust fields. New X509_STORE trust and
503 purpose functions and tidy up setting in other SSL functions.
506 +) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
507 structure. These are inherited by X509_STORE_CTX when it is
508 initialised. This allows various defaults to be set in the
509 X509_STORE structure (such as flags for CRL checking and custom
510 purpose or trust settings) for functions which only use X509_STORE_CTX
511 internally such as S/MIME.
513 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
514 trust settings if they are not set in X509_STORE. This allows X509_STORE
515 purposes and trust (in S/MIME for example) to override any set by default.
517 Add command line options for CRL checking to smime, s_client and s_server
521 +) Initial CRL based revocation checking. If the CRL checking flag(s)
522 are set then the CRL is looked up in the X509_STORE structure and
523 its validity and signature checked, then if the certificate is found
524 in the CRL the verify fails with a revoked error.
526 Various new CRL related callbacks added to X509_STORE_CTX structure.
528 Command line options added to 'verify' application to support this.
530 This needs some additional work, such as being able to handle multiple
531 CRLs with different times, extension based lookup (rather than just
532 by subject name) and ultimately more complete V2 CRL extension
536 +) Add a general user interface API (crypto/ui/). This is designed
537 to replace things like des_read_password and friends (backward
538 compatibility functions using this new API are provided).
539 The purpose is to remove prompting functions from the DES code
540 section as well as provide for prompting through dialog boxes in
541 a window system and the like.
544 *) Fix various bugs related to DSA S/MIME verification. Handle missing
545 parameters in DSA public key structures and return an error in the
546 DSA routines if parameters are absent.
549 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
550 in the current directory if neither $RANDFILE nor $HOME was set.
551 RAND_file_name() in 0.9.6a returned NULL in this case. This has
552 caused some confusion to Windows users who haven't defined $HOME.
553 Thus RAND_file_name() is changed again: e_os.h can define a
554 DEFAULT_HOME, which will be used if $HOME is not set.
555 For Windows, we use "C:"; on other platforms, we still require
556 environment variables.
558 +) Add "ex_data" support to ENGINE so implementations can add state at a
559 per-structure level rather than having to store it globally.
562 +) Make it possible for ENGINE structures to be copied when retrieved by
563 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
564 This causes the "original" ENGINE structure to act like a template,
565 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
566 operational state can be localised to each ENGINE structure, despite the
567 fact they all share the same "methods". New ENGINE structures returned in
568 this case have no functional references and the return value is the single
569 structural reference. This matches the single structural reference returned
570 by ENGINE_by_id() normally, when it is incremented on the pre-existing
574 +) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
575 needs to match any other type at all we need to manually clear the
579 +) Changes to the "openssl engine" utility to include;
580 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
581 about an ENGINE's available control commands.
582 - executing control commands from command line arguments using the
583 '-pre' and '-post' switches. '-post' is only used if '-t' is
584 specified and the ENGINE is successfully initialised. The syntax for
585 the individual commands are colon-separated, for example;
586 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
589 +) New dynamic control command support for ENGINEs. ENGINEs can now
590 declare their own commands (numbers), names (strings), descriptions,
591 and input types for run-time discovery by calling applications. A
592 subset of these commands are implicitly classed as "executable"
593 depending on their input type, and only these can be invoked through
594 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
595 can be based on user input, config files, etc). The distinction is
596 that "executable" commands cannot return anything other than a boolean
597 result and can only support numeric or string input, whereas some
598 discoverable commands may only be for direct use through
599 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
600 pointers, or other custom uses. The "executable" commands are to
601 support parameterisations of ENGINE behaviour that can be
602 unambiguously defined by ENGINEs and used consistently across any
603 OpenSSL-based application. Commands have been added to all the
604 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
605 control over shared-library paths without source code alterations.
608 +) Changed all ENGINE implementations to dynamically allocate their
609 ENGINEs rather than declaring them statically. Apart from this being
610 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
611 this also allows the implementations to compile without using the
612 internal engine_int.h header.
615 +) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
616 'const' value. Any code that should be able to modify a RAND_METHOD
617 should already have non-const pointers to it (ie. they should only
618 modify their own ones).
621 +) Made a variety of little tweaks to the ENGINE code.
622 - "atalla" and "ubsec" string definitions were moved from header files
623 to C code. "nuron" string definitions were placed in variables
624 rather than hard-coded - allowing parameterisation of these values
625 later on via ctrl() commands.
626 - Removed unused "#if 0"'d code.
627 - Fixed engine list iteration code so it uses ENGINE_free() to release
628 structural references.
629 - Constified the RAND_METHOD element of ENGINE structures.
630 - Constified various get/set functions as appropriate and added
631 missing functions (including a catch-all ENGINE_cpy that duplicates
632 all ENGINE values onto a new ENGINE except reference counts/state).
633 - Removed NULL parameter checks in get/set functions. Setting a method
634 or function to NULL is a way of cancelling out a previously set
635 value. Passing a NULL ENGINE parameter is just plain stupid anyway
636 and doesn't justify the extra error symbols and code.
637 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
638 flags from engine_int.h to engine.h.
639 - Changed prototypes for ENGINE handler functions (init(), finish(),
640 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
643 *) Move 'if (!initialized) RAND_poll()' into regions protected by
644 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
645 having multiple threads call RAND_poll() concurrently.
648 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
649 combination of a flag and a thread ID variable.
650 Otherwise while one thread is in ssleay_rand_bytes (which sets the
651 flag), *other* threads can enter ssleay_add_bytes without obeying
652 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
653 that they do not hold after the first thread unsets add_do_not_lock).
656 +) Implement binary inversion algorithm for BN_mod_inverse in addition
657 to the algorithm using long division. The binary algorithm can be
658 used only if the modulus is odd. On 32-bit systems, it is faster
659 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
660 roughly 5-15% for 256-bit moduli), so we use it only for moduli
661 up to 450 bits. In 64-bit environments, the binary algorithm
662 appears to be advantageous for much longer moduli; here we use it
663 for moduli up to 2048 bits.
666 *) Change bctest again: '-x' expressions are not available in all
670 -) OpenSSL 0.9.6a released [5 Apr 2001]
672 *) Fix a couple of memory leaks in PKCS7_dataDecode()
673 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
675 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
676 the default extension for executables, if any. Also, make the perl
677 scripts that use symlink() to test if it really exists and use "cp"
678 if it doesn't. All this made OpenSSL compilable and installable in
682 +) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
683 could not support the combine flag in choice fields.
686 -) Fix for asn1_GetSequence() for indefinite length constructed data.
687 If SEQUENCE is length is indefinite just set c->slen to the total
688 amount of data available.
689 [Steve Henson, reported by shige@FreeBSD.org]
690 [This change does not apply to 0.9.7.]
692 *) Change bctest to avoid here-documents inside command substitution
693 (workaround for FreeBSD /bin/sh bug).
694 For compatibility with Ultrix, avoid shell functions (introduced
695 in the bctest version that searches along $PATH).
698 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
699 with des_encrypt() defined on some operating systems, like Solaris
703 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
704 On the Importance of Eliminating Errors in Cryptographic
705 Computations, J. Cryptology 14 (2001) 2, 101-119,
706 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
709 *) MIPS assembler BIGNUM division bug fix.
712 *) Disabled incorrect Alpha assembler code.
715 -) Fix PKCS#7 decode routines so they correctly update the length
716 after reading an EOC for the EXPLICIT tag.
718 [This change does not apply to 0.9.7.]
720 *) Fix bug in PKCS#12 key generation routines. This was triggered
721 if a 3DES key was generated with a 0 initial byte. Include
722 PKCS12_BROKEN_KEYGEN compilation option to retain the old
723 (but broken) behaviour.
726 *) Enhance bctest to search for a working bc along $PATH and print
728 [Tim Rice <tim@multitalents.net> via Richard Levitte]
730 +) Add a 'copy_extensions' option to the 'ca' utility. This copies
731 extensions from a certificate request to the certificate.
734 +) Allow multiple 'certopt' and 'nameopt' options to be separated
735 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
736 file: this allows the display of the certificate about to be
737 signed to be customised, to allow certain fields to be included
738 or excluded and extension details. The old system didn't display
739 multicharacter strings properly, omitted fields not in the policy
740 and couldn't display additional details such as extensions.
743 *) Fix memory leaks in err.c: free err_data string if necessary;
744 don't write to the wrong index in ERR_set_error_data.
747 +) Function EC_POINTs_mul for simultaneous scalar multiplication
748 of an arbitrary number of elliptic curve points, optionally
749 including the generator defined for the EC_GROUP.
750 EC_POINT_mul is a simple wrapper function for the typical case
751 that the point list has just one item (besides the optional
755 +) First EC_METHODs for curves over GF(p):
757 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
758 operations and provides various method functions that can also
759 operate with faster implementations of modular arithmetic.
761 EC_GFp_mont_method() reuses most functions that are part of
762 EC_GFp_simple_method, but uses Montgomery arithmetic.
764 [Bodo Moeller; point addition and point doubling
765 implementation directly derived from source code provided by
766 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
768 +) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
771 Curves are EC_GROUP objects (with an optional group generator)
772 based on EC_METHODs that are built into the library.
774 Points are EC_POINT objects based on EC_GROUP objects.
776 Most of the framework would be able to handle curves over arbitrary
777 finite fields, but as there are no obvious types for fields other
778 than GF(p), some functions are limited to that for now.
781 +) Add the -HTTP option to s_server. It is similar to -WWW, but requires
782 that the file contains a complete HTTP response.
785 +) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
786 change the def and num file printf format specifier from "%-40sXXX"
787 to "%-39s XXX". The latter will always guarantee a space after the
788 field while the former will cause them to run together if the field
789 is 40 of more characters long.
792 +) Constify the cipher and digest 'method' functions and structures
793 and modify related functions to take constant EVP_MD and EVP_CIPHER
797 *) Implement ssl23_peek (analogous to ssl23_read), which previously
801 *) Replace rdtsc with _emit statements for VC++ version 5.
802 [Jeremy Cooper <jeremy@baymoo.org>]
804 +) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
805 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
808 +) Modify EVP_Digest*() routines so they now return values. Although the
809 internal software routines can never fail additional hardware versions
813 +) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
815 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
816 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
819 ERR_R_NESTED_ASN1_ERROR
821 ERR_R_MISSING_ASN1_EOS
822 were 4 .. 9, conflicting with
823 ERR_LIB_RSA (= ERR_R_RSA_LIB)
825 ERR_LIB_PEM (= ERR_R_PEM_LIB).
826 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
828 Add new error code 'ERR_R_INTERNAL_ERROR'.
831 +) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
835 +) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
836 sets the subject name for a new request or supersedes the
837 subject name in a given request. Formats that can be parsed are
838 'CN=Some Name, OU=myOU, C=IT'
840 'CN=Some Name/OU=myOU/C=IT'.
842 Add options '-batch' and '-verbose' to 'openssl req'.
843 [Massimiliano Pala <madwolf@hackmasters.net>]
845 *) Make it possible to reuse SSLv2 sessions.
848 +) Introduce the possibility to access global variables through
849 functions on platform were that's the best way to handle exporting
850 global variables in shared libraries. To enable this functionality,
851 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
852 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
853 is normally done by Configure or something similar).
855 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
856 in the source file (foo.c) like this:
858 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
859 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
861 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
862 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
864 OPENSSL_DECLARE_GLOBAL(int,foo);
865 #define foo OPENSSL_GLOBAL_REF(foo)
866 OPENSSL_DECLARE_GLOBAL(double,bar);
867 #define bar OPENSSL_GLOBAL_REF(bar)
869 The #defines are very important, and therefore so is including the
870 header file everywhere where the defined globals are used.
872 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
873 of ASN.1 items, but that structure is a bit different.
875 The largest change is in util/mkdef.pl which has been enhanced with
876 better and easier to understand logic to choose which symbols should
877 go into the Windows .def files as well as a number of fixes and code
878 cleanup (among others, algorithm keywords are now sorted
879 lexicographically to avoid constant rewrites).
882 *) In copy_email() check for >= 0 as a return value for
883 X509_NAME_get_index_by_NID() since 0 is a valid index.
884 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
886 +) In BN_div() keep a copy of the sign of 'num' before writing the
887 result to 'rm' because if rm==num the value will be overwritten
888 and produce the wrong result if 'num' is negative: this caused
889 problems with BN_mod() and BN_nnmod().
892 +) Function OCSP_request_verify(). This checks the signature on an
893 OCSP request and verifies the signer certificate. The signer
894 certificate is just checked for a generic purpose and OCSP request
898 +) Add OCSP_check_validity() function to check the validity of OCSP
899 responses. OCSP responses are prepared in real time and may only
900 be a few seconds old. Simply checking that the current time lies
901 between thisUpdate and nextUpdate max reject otherwise valid responses
902 caused by either OCSP responder or client clock inaccuracy. Instead
903 we allow thisUpdate and nextUpdate to fall within a certain period of
904 the current time. The age of the response can also optionally be
905 checked. Two new options -validity_period and -status_age added to
909 +) If signature or public key algorithm is unrecognized print out its
910 OID rather that just UNKNOWN.
913 *) Avoid coredump with unsupported or invalid public keys by checking if
914 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
915 PKCS7_verify() fails with non detached data.
918 +) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
919 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
920 ID to be generated from the issuer certificate alone which can then be
921 passed to OCSP_id_issuer_cmp().
924 +) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
925 ASN1 modules to export functions returning ASN1_ITEM pointers
926 instead of the ASN1_ITEM structures themselves. This adds several
927 new macros which allow the underlying ASN1 function/structure to
928 be accessed transparently. As a result code should not use ASN1_ITEM
929 references directly (such as &X509_it) but instead use the relevant
930 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
931 use of the new ASN1 code on platforms where exporting structures
932 is problematical (for example in shared libraries) but exporting
933 functions returning pointers to structures is not.
936 +) Add support for overriding the generation of SSL/TLS session IDs.
937 These callbacks can be registered either in an SSL_CTX or per SSL.
938 The purpose of this is to allow applications to control, if they wish,
939 the arbitrary values chosen for use as session IDs, particularly as it
940 can be useful for session caching in multiple-server environments. A
941 command-line switch for testing this (and any client code that wishes
942 to use such a feature) has been added to "s_server".
943 [Geoff Thorpe, Lutz Jaenicke]
945 +) Modify mkdef.pl to recognise and parse preprocessor conditionals
946 of the form '#if defined(...) || defined(...) || ...' and
947 '#if !defined(...) && !defined(...) && ...'. This also avoids
948 the growing number of special cases it was previously handling.
951 *) Don't use getenv in library functions when run as setuid/setgid.
952 New function OPENSSL_issetugid().
955 +) Make all configuration macros available for application by making
956 sure they are available in opensslconf.h, by giving them names starting
957 with "OPENSSL_" to avoid conflicts with other packages and by making
958 sure e_os2.h will cover all platform-specific cases together with
960 Additionally, it is now possible to define configuration/platform-
961 specific names (called "system identities"). In the C code, these
962 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
963 macro with the name beginning with "OPENSSL_SYS_", which is determined
964 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
968 +) New option -set_serial to 'req' and 'x509' this allows the serial
969 number to use to be specified on the command line. Previously self
970 signed certificates were hard coded with serial number 0 and the
971 CA options of 'x509' had to use a serial number in a file which was
975 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
976 due to incorrect handling of multi-threading:
978 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
980 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
982 3. Count how many times MemCheck_off() has been called so that
983 nested use can be treated correctly. This also avoids
984 inband-signalling in the previous code (which relied on the
985 assumption that thread ID 0 is impossible).
988 +) New options to 'ca' utility to support V2 CRL entry extensions.
989 Currently CRL reason, invalidity date and hold instruction are
990 supported. Add new CRL extensions to V3 code and some new objects.
993 *) Add "-rand" option also to s_client and s_server.
996 +) New function EVP_CIPHER_CTX_set_padding() this is used to
997 disable standard block padding (aka PKCS#5 padding) in the EVP
998 API, which was previously mandatory. This means that the data is
999 not padded in any way and so the total length much be a multiple
1000 of the block size, otherwise an error occurs.
1003 +) Initial (incomplete) OCSP SSL support.
1006 *) Fix CPU detection on Irix 6.x.
1007 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
1008 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
1010 +) New function OCSP_parse_url(). This splits up a URL into its host,
1011 port and path components: primarily to parse OCSP URLs. New -url
1012 option to ocsp utility.
1015 +) New nonce behavior. The return value of OCSP_check_nonce() now
1016 reflects the various checks performed. Applications can decide
1017 whether to tolerate certain situations such as an absent nonce
1018 in a response when one was present in a request: the ocsp application
1019 just prints out a warning. New function OCSP_add1_basic_nonce()
1020 this is to allow responders to include a nonce in a response even if
1021 the request is nonce-less.
1024 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
1027 [This change does not apply to 0.9.7.]
1029 *) Use the cached encoding of an X509_NAME structure rather than
1030 copying it. This is apparently the reason for the libsafe "errors"
1031 but the code is actually correct.
1034 +) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
1035 skipped when using openssl x509 multiple times on a single input file,
1036 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
1039 +) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
1040 set string type: to handle setting ASN1_TIME structures. Fix ca
1041 utility to correctly initialize revocation date of CRLs.
1044 +) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
1045 the clients preferred ciphersuites and rather use its own preferences.
1046 Should help to work around M$ SGC (Server Gated Cryptography) bug in
1047 Internet Explorer by ensuring unchanged hash method during stepup.
1048 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
1051 +) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
1052 to aes and add a new 'exist' option to print out symbols that don't
1056 +) Additional options to ocsp utility to allow flags to be set and
1057 additional certificates supplied.
1060 +) Add the option -VAfile to 'openssl ocsp', so the user can give the
1061 OCSP client a number of certificate to only verify the response
1065 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
1066 Bleichenbacher's DSA attack.
1067 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
1068 to be set and top=0 forces the highest bit to be set; top=-1 is new
1069 and leaves the highest bit random.
1070 [Ulf Moeller, Bodo Moeller]
1072 +) Update Rijndael code to version 3.0 and change EVP AES ciphers to
1073 handle the new API. Currently only ECB, CBC modes supported. Add new
1074 AES OIDs. Add TLS AES ciphersuites as described in the "AES Ciphersuites
1075 for TLS" draft-ietf-tls-ciphersuite-03.txt.
1076 [Ben Laurie, Steve Henson]
1078 *) In the NCONF_...-based implementations for CONF_... queries
1079 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
1080 a temporary CONF structure with the data component set to NULL
1081 (which gives segmentation faults in lh_retrieve).
1082 Instead, use NULL for the CONF pointer in CONF_get_string and
1083 CONF_get_number (which may use environment variables) and directly
1084 return NULL from CONF_get_section.
1087 *) Fix potential buffer overrun for EBCDIC.
1090 +) New function OCSP_copy_nonce() to copy nonce value (if present) from
1091 request to response.
1094 +) Functions for OCSP responders. OCSP_request_onereq_count(),
1095 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
1096 extract information from a certificate request. OCSP_response_create()
1097 creates a response and optionally adds a basic response structure.
1098 OCSP_basic_add1_status() adds a complete single response to a basic
1099 response and returns the OCSP_SINGLERESP structure just added (to allow
1100 extensions to be included for example). OCSP_basic_add1_cert() adds a
1101 certificate to a basic response and OCSP_basic_sign() signs a basic
1102 response with various flags. New helper functions ASN1_TIME_check()
1103 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
1104 (converts ASN1_TIME to GeneralizedTime).
1107 +) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
1108 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
1109 structure from a certificate. X509_pubkey_digest() digests the public_key
1110 contents: this is used in various key identifiers.
1113 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
1114 keyUsage if basicConstraints absent for a CA.
1117 *) Make SMIME_write_PKCS7() write mail header values with a format that
1118 is more generally accepted (no spaces before the semicolon), since
1119 some programs can't parse those values properly otherwise. Also make
1120 sure BIO's that break lines after each write do not create invalid
1124 +) Make sk_sort() tolerate a NULL argument.
1125 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
1127 +) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
1128 passed by the function are trusted implicitly. If any of them signed the
1129 response then it is assumed to be valid and is not verified.
1132 -) Make the CRL encoding routines work with empty SEQUENCE OF. The
1133 macros previously used would not encode an empty SEQUENCE OF
1134 and break the signature.
1136 [This change does not apply to 0.9.7.]
1138 *) Zero the premaster secret after deriving the master secret in
1142 +) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
1143 to data. This was previously part of the PKCS7 ASN1 code. This
1144 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
1145 [Steve Henson, reported by Kenneth R. Robinette
1146 <support@securenetterm.com>]
1148 +) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
1149 routines: without these tracing memory leaks is very painful.
1150 Fix leaks in PKCS12 and PKCS7 routines.
1153 *) Add some EVP_add_digest_alias registrations (as found in
1154 OpenSSL_add_all_digests()) to SSL_library_init()
1155 aka OpenSSL_add_ssl_algorithms(). This provides improved
1156 compatibility with peers using X.509 certificates
1157 with unconventional AlgorithmIdentifier OIDs.
1160 *) Fix for Irix with NO_ASM.
1161 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
1163 *) ./config script fixes.
1164 [Ulf Moeller, Richard Levitte]
1166 +) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
1167 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
1168 effectively meant GeneralizedTime would never be used. Now it
1169 is initialised to -1 but X509_time_adj() now has to check the value
1170 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
1171 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
1172 [Steve Henson, reported by Kenneth R. Robinette
1173 <support@securenetterm.com>]
1175 +) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
1176 result in a zero length in the ASN1_INTEGER structure which was
1177 not consistent with the structure when d2i_ASN1_INTEGER() was used
1178 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
1179 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
1180 where it did not print out a minus for negative ASN1_INTEGER.
1183 *) Fix 'openssl passwd -1'.
1186 +) Add summary printout to ocsp utility. The various functions which
1187 convert status values to strings have been renamed to:
1188 OCSP_response_status_str(), OCSP_cert_status_str() and
1189 OCSP_crl_reason_str() and are no longer static. New options
1190 to verify nonce values and to disable verification. OCSP response
1191 printout format cleaned up.
1194 +) Add additional OCSP certificate checks. These are those specified
1195 in RFC2560. This consists of two separate checks: the CA of the
1196 certificate being checked must either be the OCSP signer certificate
1197 or the issuer of the OCSP signer certificate. In the latter case the
1198 OCSP signer certificate must contain the OCSP signing extended key
1199 usage. This check is performed by attempting to match the OCSP
1200 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
1201 in the OCSP_CERTID structures of the response.
1204 +) Initial OCSP certificate verification added to OCSP_basic_verify()
1205 and related routines. This uses the standard OpenSSL certificate
1206 verify routines to perform initial checks (just CA validity) and
1207 to obtain the certificate chain. Then additional checks will be
1208 performed on the chain. Currently the root CA is checked to see
1209 if it is explicitly trusted for OCSP signing. This is used to set
1210 a root CA as a global signing root: that is any certificate that
1211 chains to that CA is an acceptable OCSP signing certificate.
1214 +) New '-extfile ...' option to 'openssl ca' for reading X.509v3
1215 extensions from a separate configuration file.
1216 As when reading extensions from the main configuration file,
1217 the '-extensions ...' option may be used for specifying the
1219 [Massimiliano Pala <madwolf@comune.modena.it>]
1221 *) Change PKCS12_key_gen_asc() so it can cope with non null
1222 terminated strings whose length is passed in the passlen
1223 parameter, for example from PEM callbacks. This was done
1224 by adding an extra length parameter to asc2uni().
1225 [Steve Henson, reported by <oddissey@samsung.co.kr>]
1227 +) New OCSP utility. Allows OCSP requests to be generated or
1228 read. The request can be sent to a responder and the output
1229 parsed, outputed or printed in text form. Not complete yet:
1230 still needs to check the OCSP response validity.
1233 +) New subcommands for 'openssl ca':
1234 'openssl ca -status <serial>' prints the status of the cert with
1235 the given serial number (according to the index file).
1236 'openssl ca -updatedb' updates the expiry status of certificates
1238 [Massimiliano Pala <madwolf@comune.modena.it>]
1240 +) New '-newreq-nodes' command option to CA.pl. This is like
1241 '-newreq', but calls 'openssl req' with the '-nodes' option
1242 so that the resulting key is not encrypted.
1243 [Damien Miller <djm@mindrot.org>]
1245 +) New configuration for the GNU Hurd.
1246 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
1248 +) Initial code to implement OCSP basic response verify. This
1249 is currently incomplete. Currently just finds the signer's
1250 certificate and verifies the signature on the response.
1253 +) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
1254 value of OPENSSLDIR. This is available via the new '-d' option
1255 to 'openssl version', and is also included in 'openssl version -a'.
1258 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
1259 call failed, free the DSA structure.
1262 +) Allowing defining memory allocation callbacks that will be given
1263 file name and line number information in additional arguments
1264 (a const char* and an int). The basic functionality remains, as
1265 well as the original possibility to just replace malloc(),
1266 realloc() and free() by functions that do not know about these
1267 additional arguments. To register and find out the current
1268 settings for extended allocation functions, the following
1269 functions are provided:
1271 CRYPTO_set_mem_ex_functions
1272 CRYPTO_set_locked_mem_ex_functions
1273 CRYPTO_get_mem_ex_functions
1274 CRYPTO_get_locked_mem_ex_functions
1276 These work the same way as CRYPTO_set_mem_functions and friends.
1277 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
1278 extended allocation function is enabled.
1279 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
1280 a conventional allocation function is enabled.
1281 [Richard Levitte, Bodo Moeller]
1283 *) Fix to uni2asc() to cope with zero length Unicode strings.
1284 These are present in some PKCS#12 files.
1287 +) Finish off removing the remaining LHASH function pointer casts.
1288 There should no longer be any prototype-casting required when using
1289 the LHASH abstraction, and any casts that remain are "bugs". See
1290 the callback types and macros at the head of lhash.h for details
1291 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
1294 +) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
1295 If /dev/[u]random devices are not available or do not return enough
1296 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
1298 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
1299 /etc/entropy will be queried once each in this sequence, quering stops
1300 when enough entropy was collected without querying more sockets.
1303 +) Change the Unix RAND_poll() variant to be able to poll several
1304 random devices, as specified by DEVRANDOM, until a sufficient amount
1305 of data has been collected. We spend at most 10 ms on each file
1306 (select timeout) and read in non-blocking mode. DEVRANDOM now
1307 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
1308 (previously it was just the string "/dev/urandom"), so on typical
1309 platforms the 10 ms delay will never occur.
1310 Also separate out the Unix variant to its own file, rand_unix.c.
1311 For VMS, there's a currently-empty rand_vms.c.
1314 +) Move OCSP client related routines to ocsp_cl.c. These
1315 provide utility functions which an application needing
1316 to issue a request to an OCSP responder and analyse the
1317 response will typically need: as opposed to those which an
1318 OCSP responder itself would need which will be added later.
1320 OCSP_request_sign() signs an OCSP request with an API similar
1321 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
1322 response. OCSP_response_get1_basic() extracts basic response
1323 from response. OCSP_resp_find_status(): finds and extracts status
1324 information from an OCSP_CERTID structure (which will be created
1325 when the request structure is built). These are built from lower
1326 level functions which work on OCSP_SINGLERESP structures but
1327 wont normally be used unless the application wishes to examine
1328 extensions in the OCSP response for example.
1330 Replace nonce routines with a pair of functions.
1331 OCSP_request_add1_nonce() adds a nonce value and optionally
1332 generates a random value. OCSP_check_nonce() checks the
1333 validity of the nonce in an OCSP response.
1336 +) Change function OCSP_request_add() to OCSP_request_add0_id().
1337 This doesn't copy the supplied OCSP_CERTID and avoids the
1338 need to free up the newly created id. Change return type
1339 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
1340 This can then be used to add extensions to the request.
1341 Deleted OCSP_request_new(), since most of its functionality
1342 is now in OCSP_REQUEST_new() (and the case insensitive name
1343 clash) apart from the ability to set the request name which
1344 will be added elsewhere.
1347 +) Update OCSP API. Remove obsolete extensions argument from
1348 various functions. Extensions are now handled using the new
1349 OCSP extension code. New simple OCSP HTTP function which
1350 can be used to send requests and parse the response.
1353 +) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
1354 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
1355 uses the special reorder version of SET OF to sort the attributes
1356 and reorder them to match the encoded order. This resolves a long
1357 standing problem: a verify on a PKCS7 structure just after signing
1358 it used to fail because the attribute order did not match the
1359 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
1360 it uses the received order. This is necessary to tolerate some broken
1361 software that does not order SET OF. This is handled by encoding
1362 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
1363 to produce the required SET OF.
1366 +) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
1367 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
1368 files to get correct declarations of the ASN.1 item variables.
1371 +) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
1372 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
1373 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
1374 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
1375 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
1376 ASN1_ITEM and no wrapper functions.
1379 +) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
1380 replace the old function pointer based I/O routines. Change most of
1381 the *_d2i_bio() and *_d2i_fp() functions to use these.
1384 +) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
1385 lines, recognice more "algorithms" that can be deselected, and make
1386 it complain about algorithm deselection that isn't recognised.
1389 +) New ASN1 functions to handle dup, sign, verify, digest, pack and
1390 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
1391 to use new functions. Add NO_ASN1_OLD which can be set to remove
1392 some old style ASN1 functions: this can be used to determine if old
1393 code will still work when these eventually go away.
1396 +) New extension functions for OCSP structures, these follow the
1397 same conventions as certificates and CRLs.
1400 +) New function X509V3_add1_i2d(). This automatically encodes and
1401 adds an extension. Its behaviour can be customised with various
1402 flags to append, replace or delete. Various wrappers added for
1403 certifcates and CRLs.
1406 +) Fix to avoid calling the underlying ASN1 print routine when
1407 an extension cannot be parsed. Correct a typo in the
1408 OCSP_SERVICELOC extension. Tidy up print OCSP format.
1411 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
1412 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
1413 when writing a 32767 byte record.
1414 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
1416 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
1417 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
1419 (RSA objects have a reference count access to which is protected
1420 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
1421 so they are meant to be shared between threads.)
1422 [Bodo Moeller, Geoff Thorpe; original patch submitted by
1423 "Reddie, Steven" <Steven.Reddie@ca.com>]
1425 +) Make mkdef.pl parse some of the ASN1 macros and add apropriate
1426 entries for variables.
1429 *) Fix a deadlock in CRYPTO_mem_leaks().
1432 +) Add functionality to apps/openssl.c for detecting locking
1433 problems: As the program is single-threaded, all we have
1434 to do is register a locking callback using an array for
1435 storing which locks are currently held by the program.
1438 +) Use a lock around the call to CRYPTO_get_ex_new_index() in
1439 SSL_get_ex_data_X509_STORE_idx(), which is used in
1440 ssl_verify_cert_chain() and thus can be called at any time
1441 during TLS/SSL handshakes so that thread-safety is essential.
1442 Unfortunately, the ex_data design is not at all suited
1443 for multi-threaded use, so it probably should be abolished.
1446 +) Added Broadcom "ubsec" ENGINE to OpenSSL.
1447 [Broadcom, tweaked and integrated by Geoff Thorpe]
1449 +) Move common extension printing code to new function
1450 X509V3_print_extensions(). Reorganise OCSP print routines and
1451 implement some needed OCSP ASN1 functions. Add OCSP extensions.
1454 +) New function X509_signature_print() to remove duplication in some
1458 +) Add a special meaning when SET OF and SEQUENCE OF flags are both
1459 set (this was treated exactly the same as SET OF previously). This
1460 is used to reorder the STACK representing the structure to match the
1461 encoding. This will be used to get round a problem where a PKCS7
1462 structure which was signed could not be verified because the STACK
1463 order did not reflect the encoded order.
1466 +) Reimplement the OCSP ASN1 module using the new code.
1469 +) Update the X509V3 code to permit the use of an ASN1_ITEM structure
1470 for its ASN1 operations. The old style function pointers still exist
1471 for now but they will eventually go away.
1474 +) Merge in replacement ASN1 code from the ASN1 branch. This almost
1475 completely replaces the old ASN1 functionality with a table driven
1476 encoder and decoder which interprets an ASN1_ITEM structure describing
1477 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
1478 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
1479 has also been converted to the new form.
1482 +) Change BN_mod_exp_recp so that negative moduli are tolerated
1483 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
1484 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
1485 for negative moduli.
1488 +) Fix BN_uadd and BN_usub: Always return non-negative results instead
1489 of not touching the result's sign bit.
1492 +) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
1496 +) Changed the LHASH code to use prototypes for callbacks, and created
1497 macros to declare and implement thin (optionally static) functions
1498 that provide type-safety and avoid function pointer casting for the
1499 type-specific callbacks.
1502 *) Use better test patterns in bntest.
1505 +) Added Kerberos Cipher Suites to be used with TLS, as written in
1507 [Veers Staats <staatsvr@asc.hpc.mil>,
1508 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
1510 *) rand_win.c fix for Borland C.
1513 *) BN_rshift bugfix for n == 0.
1516 +) Reformat the FAQ so the different questions and answers can be divided
1517 in sections depending on the subject.
1520 +) Have the zlib compression code load ZLIB.DLL dynamically under
1524 +) New function BN_mod_sqrt for computing square roots modulo a prime
1525 (using the probabilistic Tonelli-Shanks algorithm unless
1526 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
1527 be handled deterministically).
1528 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
1530 *) Add a 'bctest' script that checks for some known 'bc' bugs
1531 so that 'make test' does not abort just because 'bc' is broken.
1534 *) Store verify_result within SSL_SESSION also for client side to
1535 avoid potential security hole. (Re-used sessions on the client side
1536 always resulted in verify_result==X509_V_OK, not using the original
1537 result of the server certificate verification.)
1540 +) Make BN_mod_inverse faster by explicitly handling small quotients
1541 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
1542 512 bits], about 30% for larger ones [1024 or 2048 bits].)
1545 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
1546 SSL3_RT_APPLICATION_DATA, return 0.
1547 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
1551 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
1552 releases, have been re-implemented by renaming the previous
1553 implementations of ssl2_read and ssl3_read to ssl2_read_internal
1554 and ssl3_read_internal, respectively, and adding 'peek' parameters
1555 to them. The new ssl[23]_{read,peek} functions are calls to
1556 ssl[23]_read_internal with the 'peek' flag set appropriately.
1557 A 'peek' parameter has also been added to ssl3_read_bytes, which
1558 does the actual work for ssl3_read_internal.
1561 +) New function BN_kronecker.
1564 +) Fix BN_gcd so that it works on negative inputs; the result is
1565 positive unless both parameters are zero.
1566 Previously something reasonably close to an infinite loop was
1567 possible because numbers could be growing instead of shrinking
1568 in the implementation of Euclid's algorithm.
1571 +) Fix BN_is_word() and BN_is_one() macros to take into account the
1572 sign of the number in question.
1574 Fix BN_is_word(a,w) to work correctly for w == 0.
1576 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
1577 because its test if the absolute value of 'a' equals 'w'.
1578 Note that BN_abs_is_word does *not* handle w == 0 reliably;
1579 it exists mostly for use in the implementations of BN_is_zero(),
1580 BN_is_one(), and BN_is_word().
1583 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
1584 the method-specific "init()" handler. Also clean up ex_data after
1585 calling the method-specific "finish()" handler. Previously, this was
1586 happening the other way round.
1589 +) New function BN_swap.
1592 +) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
1593 the exponentiation functions are more likely to produce reasonable
1594 results on negative inputs.
1597 +) Change BN_mod_mul so that the result is always non-negative.
1598 Previously, it could be negative if one of the factors was negative;
1599 I don't think anyone really wanted that behaviour.
1602 +) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
1603 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
1604 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
1605 and add new functions:
1614 BN_mod_lshift1_quick
1618 These functions always generate non-negative results.
1620 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
1621 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
1623 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
1624 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
1625 be reduced modulo m.
1626 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
1628 +) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
1629 was actually never needed) and in BN_mul(). The removal in BN_mul()
1630 required a small change in bn_mul_part_recursive() and the addition
1631 of the functions bn_cmp_part_words(), bn_sub_part_words() and
1632 bn_add_part_words(), which do the same thing as bn_cmp_words(),
1633 bn_sub_words() and bn_add_words() except they take arrays with
1637 +) In 'openssl passwd', verify passwords read from the terminal
1638 unless the '-salt' option is used (which usually means that
1639 verification would just waste user's time since the resulting
1640 hash is going to be compared with some given password hash)
1641 or the new '-noverify' option is used.
1643 This is an incompatible change, but it does not affect
1644 non-interactive use of 'openssl passwd' (passwords on the command
1645 line, '-stdin' option, '-in ...' option) and thus should not
1649 +) Remove all references to RSAref, since there's no more need for it.
1652 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
1653 The previous value, 12, was not always sufficient for BN_mod_exp().
1656 +) Make DSO load along a path given through an environment variable
1657 (SHLIB_PATH) with shl_load().
1660 +) Constify the ENGINE code as a result of BIGNUM constification.
1661 Also constify the RSA code and most things related to it. In a
1662 few places, most notable in the depth of the ASN.1 code, ugly
1663 casts back to non-const were required (to be solved at a later
1667 +) Make it so the openssl application has all engines loaded by default.
1670 +) Constify the BIGNUM routines a little more.
1673 *) Make sure that shared libraries get the internal name engine with
1674 the full version number and not just 0. This should mark the
1675 shared libraries as not backward compatible. Of course, this should
1676 be changed again when we can guarantee backward binary compatibility.
1679 +) Add the following functions:
1681 ENGINE_load_cswift()
1683 ENGINE_load_atalla()
1685 ENGINE_load_builtin_engines()
1687 That way, an application can itself choose if external engines that
1688 are built-in in OpenSSL shall ever be used or not. The benefit is
1689 that applications won't have to be linked with libdl or other dso
1690 libraries unless it's really needed.
1692 Changed 'openssl engine' to load all engines on demand.
1693 Changed the engine header files to avoid the duplication of some
1694 declarations (they differed!).
1697 +) 'openssl engine' can now list capabilities.
1700 +) Better error reporting in 'openssl engine'.
1703 +) Never call load_dh_param(NULL) in s_server.
1706 +) Add engine application. It can currently list engines by name and
1707 identity, and test if they are actually available.
1710 +) Add support for shared libraries under Irix.
1711 [Albert Chin-A-Young <china@thewrittenword.com>]
1713 +) Improve RPM specification file by forcing symbolic linking and making
1714 sure the installed documentation is also owned by root.root.
1715 [Damien Miller <djm@mindrot.org>]
1717 +) Add configuration option to build on Linux on both big-endian and
1719 [Ralf Baechle <ralf@uni-koblenz.de>]
1721 +) Give the OpenSSL applications more possibilities to make use of
1722 keys (public as well as private) handled by engines.
1725 +) Add OCSP code that comes from CertCo.
1728 +) Add VMS support for the Rijndael code.
1731 +) Added untested support for Nuron crypto accelerator.
1734 +) Add support for external cryptographic devices. This code was
1735 previously distributed separately as the "engine" branch.
1736 [Geoff Thorpe, Richard Levitte]
1738 +) Rework the filename-translation in the DSO code. It is now possible to
1739 have far greater control over how a "name" is turned into a filename
1740 depending on the operating environment and any oddities about the
1741 different shared library filenames on each system.
1744 +) Support threads on FreeBSD-elf in Configure.
1747 +) Add the possibility to create shared libraries on HP-UX
1750 +) Fix for SHA1 assembly problem with MASM: it produces
1751 warnings about corrupt line number information when assembling
1752 with debugging information. This is caused by the overlapping
1754 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
1757 NCONF_get_number() has no error checking at all. As a replacement,
1758 NCONF_get_number_e() is defined (_e for "error checking") and is
1759 promoted strongly. The old NCONF_get_number is kept around for
1760 binary backward compatibility.
1761 Make it possible for methods to load from something other than a BIO,
1762 by providing a function pointer that is given a name instead of a BIO.
1763 For example, this could be used to load configuration data from an
1767 *) Fix typo in get_cert_by_subject() in by_dir.c
1768 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
1770 *) Rework the system to generate shared libraries:
1772 - Make note of the expected extension for the shared libraries and
1773 if there is a need for symbolic links from for example libcrypto.so.0
1774 to libcrypto.so.0.9.7. There is extended info in Configure for
1777 - Make as few rebuilds of the shared libraries as possible.
1779 - Still avoid linking the OpenSSL programs with the shared libraries.
1781 - When installing, install the shared libraries separately from the
1785 +) Fix for non blocking accept BIOs. Added new I/O special reason
1786 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
1787 with non blocking I/O was not possible because no retry code was
1788 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
1792 +) Added the beginnings of Rijndael support.
1795 +) Fix for bug in DirectoryString mask setting. Add support for
1796 X509_NAME_print_ex() in 'req' and X509_print_ex() function
1797 to allow certificate printing to more controllable, additional
1798 'certopt' option to 'x509' to allow new printing options to be
1802 +) Clean old EAY MD5 hack from e_os.h.
1805 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
1807 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
1808 and not in SSL_clear because the latter is also used by the
1809 accept/connect functions; previously, the settings made by
1810 SSL_set_read_ahead would be lost during the handshake.
1811 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
1813 *) Correct util/mkdef.pl to be selective about disabled algorithms.
1814 Previously, it would create entries for disableed algorithms no
1818 *) Added several new manual pages for SSL_* function.
1821 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
1823 *) In ssl23_get_client_hello, generate an error message when faced
1824 with an initial SSL 3.0/TLS record that is too small to contain the
1825 first two bytes of the ClientHello message, i.e. client_version.
1826 (Note that this is a pathologic case that probably has never happened
1827 in real life.) The previous approach was to use the version number
1828 from the record header as a substitute; but our protocol choice
1829 should not depend on that one because it is not authenticated
1830 by the Finished messages.
1833 *) More robust randomness gathering functions for Windows.
1834 [Jeffrey Altman <jaltman@columbia.edu>]
1836 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
1837 not set then we don't setup the error code for issuer check errors
1838 to avoid possibly overwriting other errors which the callback does
1839 handle. If an application does set the flag then we assume it knows
1840 what it is doing and can handle the new informational codes
1844 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
1845 a general "ANY" type, as such it should be able to decode anything
1846 including tagged types. However it didn't check the class so it would
1847 wrongly interpret tagged types in the same way as their universal
1848 counterpart and unknown types were just rejected. Changed so that the
1849 tagged and unknown types are handled in the same way as a SEQUENCE:
1850 that is the encoding is stored intact. There is also a new type
1851 "V_ASN1_OTHER" which is used when the class is not universal, in this
1852 case we have no idea what the actual type is so we just lump them all
1856 *) On VMS, stdout may very well lead to a file that is written to
1857 in a record-oriented fashion. That means that every write() will
1858 write a separate record, which will be read separately by the
1859 programs trying to read from it. This can be very confusing.
1861 The solution is to put a BIO filter in the way that will buffer
1862 text until a linefeed is reached, and then write everything a
1863 line at a time, so every record written will be an actual line,
1864 not chunks of lines and not (usually doesn't happen, but I've
1865 seen it once) several lines in one record. BIO_f_linebuffer() is
1868 Currently, it's a VMS-only method, because that's where it has
1869 been tested well enough.
1872 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
1873 it can return incorrect results.
1874 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
1875 but it was in 0.9.6-beta[12].)
1878 *) Disable the check for content being present when verifying detached
1879 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
1880 include zero length content when signing messages.
1883 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
1884 BIO_ctrl (for BIO pairs).
1887 *) Add DSO method for VMS.
1890 *) Bug fix: Montgomery multiplication could produce results with the
1894 *) Add RPM specification openssl.spec and modify it to build three
1895 packages. The default package contains applications, application
1896 documentation and run-time libraries. The devel package contains
1897 include files, static libraries and function documentation. The
1898 doc package contains the contents of the doc directory. The original
1899 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
1902 *) Add a large number of documentation files for many SSL routines.
1903 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
1905 *) Add a configuration entry for Sony News 4.
1906 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
1908 *) Don't set the two most significant bits to one when generating a
1909 random number < q in the DSA library.
1912 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
1913 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
1914 the underlying transport is blocking) if a handshake took place.
1915 (The default behaviour is needed by applications such as s_client
1916 and s_server that use select() to determine when to use SSL_read;
1917 but for applications that know in advance when to expect data, it
1918 just makes things more complicated.)
1921 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
1925 *) Add a few more EBCDIC conditionals that make `req' and `x509'
1926 work better on such systems.
1927 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
1929 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
1930 Update PKCS12_parse() so it copies the friendlyName and the
1931 keyid to the certificates aux info.
1934 *) Fix bug in PKCS7_verify() which caused an infinite loop
1935 if there was more than one signature.
1936 [Sven Uszpelkat <su@celocom.de>]
1938 *) Major change in util/mkdef.pl to include extra information
1939 about each symbol, as well as presentig variables as well
1940 as functions. This change means that there's n more need
1941 to rebuild the .num files when some algorithms are excluded.
1944 *) Allow the verify time to be set by an application,
1945 rather than always using the current time.
1948 *) Phase 2 verify code reorganisation. The certificate
1949 verify code now looks up an issuer certificate by a
1950 number of criteria: subject name, authority key id
1951 and key usage. It also verifies self signed certificates
1952 by the same criteria. The main comparison function is
1953 X509_check_issued() which performs these checks.
1955 Lot of changes were necessary in order to support this
1956 without completely rewriting the lookup code.
1958 Authority and subject key identifier are now cached.
1960 The LHASH 'certs' is X509_STORE has now been replaced
1961 by a STACK_OF(X509_OBJECT). This is mainly because an
1962 LHASH can't store or retrieve multiple objects with
1963 the same hash value.
1965 As a result various functions (which were all internal
1966 use only) have changed to handle the new X509_STORE
1967 structure. This will break anything that messed round
1968 with X509_STORE internally.
1970 The functions X509_STORE_add_cert() now checks for an
1971 exact match, rather than just subject name.
1973 The X509_STORE API doesn't directly support the retrieval
1974 of multiple certificates matching a given criteria, however
1975 this can be worked round by performing a lookup first
1976 (which will fill the cache with candidate certificates)
1977 and then examining the cache for matches. This is probably
1978 the best we can do without throwing out X509_LOOKUP
1979 entirely (maybe later...).
1981 The X509_VERIFY_CTX structure has been enhanced considerably.
1983 All certificate lookup operations now go via a get_issuer()
1984 callback. Although this currently uses an X509_STORE it
1985 can be replaced by custom lookups. This is a simple way
1986 to bypass the X509_STORE hackery necessary to make this
1987 work and makes it possible to use more efficient techniques
1988 in future. A very simple version which uses a simple
1989 STACK for its trusted certificate store is also provided
1990 using X509_STORE_CTX_trusted_stack().
1992 The verify_cb() and verify() callbacks now have equivalents
1993 in the X509_STORE_CTX structure.
1995 X509_STORE_CTX also has a 'flags' field which can be used
1996 to customise the verify behaviour.
1999 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
2000 excludes S/MIME capabilities.
2003 *) When a certificate request is read in keep a copy of the
2004 original encoding of the signed data and use it when outputing
2005 again. Signatures then use the original encoding rather than
2006 a decoded, encoded version which may cause problems if the
2007 request is improperly encoded.
2010 *) For consistency with other BIO_puts implementations, call
2011 buffer_write(b, ...) directly in buffer_puts instead of calling
2014 In BIO_puts, increment b->num_write as in BIO_write.
2015 [Peter.Sylvester@EdelWeb.fr]
2017 *) Fix BN_mul_word for the case where the word is 0. (We have to use
2018 BN_zero, we may not return a BIGNUM with an array consisting of
2022 *) Avoid calling abort() from within the library when problems are
2023 detected, except if preprocessor symbols have been defined
2024 (such as REF_CHECK, BN_DEBUG etc.).
2027 *) New openssl application 'rsautl'. This utility can be
2028 used for low level RSA operations. DER public key
2029 BIO/fp routines also added.
2032 *) New Configure entry and patches for compiling on QNX 4.
2033 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
2035 *) A demo state-machine implementation was sponsored by
2036 Nuron (http://www.nuron.com/) and is now available in
2037 demos/state_machine.
2040 *) New options added to the 'dgst' utility for signature
2041 generation and verification.
2044 *) Unrecognized PKCS#7 content types are now handled via a
2045 catch all ASN1_TYPE structure. This allows unsupported
2046 types to be stored as a "blob" and an application can
2047 encode and decode it manually.
2050 *) Fix various signed/unsigned issues to make a_strex.c
2052 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
2054 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
2055 length if passed a buffer. ASN1_INTEGER_to_BN failed
2056 if passed a NULL BN and its argument was negative.
2057 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
2059 *) Modification to PKCS#7 encoding routines to output definite
2060 length encoding. Since currently the whole structures are in
2061 memory there's not real point in using indefinite length
2062 constructed encoding. However if OpenSSL is compiled with
2063 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
2066 *) Added BIO_vprintf() and BIO_vsnprintf().
2069 *) Added more prefixes to parse for in the the strings written
2070 through a logging bio, to cover all the levels that are available
2071 through syslog. The prefixes are now:
2073 PANIC, EMERG, EMR => LOG_EMERG
2074 ALERT, ALR => LOG_ALERT
2075 CRIT, CRI => LOG_CRIT
2076 ERROR, ERR => LOG_ERR
2077 WARNING, WARN, WAR => LOG_WARNING
2078 NOTICE, NOTE, NOT => LOG_NOTICE
2079 INFO, INF => LOG_INFO
2080 DEBUG, DBG => LOG_DEBUG
2082 and as before, if none of those prefixes are present at the
2083 beginning of the string, LOG_ERR is chosen.
2085 On Win32, the LOG_* levels are mapped according to this:
2087 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
2088 LOG_WARNING => EVENTLOG_WARNING_TYPE
2089 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
2093 *) Made it possible to reconfigure with just the configuration
2094 argument "reconf" or "reconfigure". The command line arguments
2095 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
2096 and are retrieved from there when reconfiguring.
2100 [Assar Westerlund <assar@sics.se>, Richard Levitte]
2102 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
2105 *) The obj_dat.pl script was messing up the sorting of object
2106 names. The reason was that it compared the quoted version
2107 of strings as a result "OCSP" > "OCSP Signing" because
2108 " > SPACE. Changed script to store unquoted versions of
2109 names and add quotes on output. It was also omitting some
2110 names from the lookup table if they were given a default
2111 value (that is if SN is missing it is given the same
2112 value as LN and vice versa), these are now added on the
2113 grounds that if an object has a name we should be able to
2114 look it up. Finally added warning output when duplicate
2115 short or long names are found.
2118 *) Changes needed for Tandem NSK.
2119 [Scott Uroff <scott@xypro.com>]
2121 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
2122 RSA_padding_check_SSLv23(), special padding was never detected
2123 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
2124 version rollback attacks was not effective.
2126 In s23_clnt.c, don't use special rollback-attack detection padding
2127 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
2128 client; similarly, in s23_srvr.c, don't do the rollback check if
2129 SSL 2.0 is the only protocol enabled in the server.
2132 *) Make it possible to get hexdumps of unprintable data with 'openssl
2133 asn1parse'. By implication, the functions ASN1_parse_dump() and
2134 BIO_dump_indent() are added.
2137 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
2138 these print out strings and name structures based on various
2139 flags including RFC2253 support and proper handling of
2140 multibyte characters. Added options to the 'x509' utility
2141 to allow the various flags to be set.
2144 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
2145 Also change the functions X509_cmp_current_time() and
2146 X509_gmtime_adj() work with an ASN1_TIME structure,
2147 this will enable certificates using GeneralizedTime in validity
2148 dates to be checked.
2151 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
2152 negative public key encodings) on by default,
2153 NO_NEG_PUBKEY_BUG can be set to disable it.
2156 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
2157 content octets. An i2c_ASN1_OBJECT is unnecessary because
2158 the encoding can be trivially obtained from the structure.
2161 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
2162 not read locks (CRYPTO_r_[un]lock).
2165 *) A first attempt at creating official support for shared
2166 libraries through configuration. I've kept it so the
2167 default is static libraries only, and the OpenSSL programs
2168 are always statically linked for now, but there are
2169 preparations for dynamic linking in place.
2170 This has been tested on Linux and Tru64.
2173 *) Randomness polling function for Win9x, as described in:
2174 Peter Gutmann, Software Generation of Practically Strong
2178 *) Fix so PRNG is seeded in req if using an already existing
2182 *) New options to smime application. -inform and -outform
2183 allow alternative formats for the S/MIME message including
2184 PEM and DER. The -content option allows the content to be
2185 specified separately. This should allow things like Netscape
2186 form signing output easier to verify.
2189 *) Fix the ASN1 encoding of tags using the 'long form'.
2192 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
2193 STRING types. These convert content octets to and from the
2194 underlying type. The actual tag and length octets are
2195 already assumed to have been read in and checked. These
2196 are needed because all other string types have virtually
2197 identical handling apart from the tag. By having versions
2198 of the ASN1 functions that just operate on content octets
2199 IMPLICIT tagging can be handled properly. It also allows
2200 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
2201 and ASN1_INTEGER are identical apart from the tag.
2204 *) Change the handling of OID objects as follows:
2206 - New object identifiers are inserted in objects.txt, following
2207 the syntax given in objects.README.
2208 - objects.pl is used to process obj_mac.num and create a new
2210 - obj_dat.pl is used to create a new obj_dat.h, using the data in
2213 This is currently kind of a hack, and the perl code in objects.pl
2214 isn't very elegant, but it works as I intended. The simplest way
2215 to check that it worked correctly is to look in obj_dat.h and
2216 check the array nid_objs and make sure the objects haven't moved
2217 around (this is important!). Additions are OK, as well as
2218 consistent name changes.
2221 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
2224 *) Addition of the command line parameter '-rand file' to 'openssl req'.
2225 The given file adds to whatever has already been seeded into the
2226 random pool through the RANDFILE configuration file option or
2227 environment variable, or the default random state file.
2230 *) mkstack.pl now sorts each macro group into lexical order.
2231 Previously the output order depended on the order the files
2232 appeared in the directory, resulting in needless rewriting
2236 *) Patches to make OpenSSL compile under Win32 again. Mostly
2237 work arounds for the VC++ problem that it treats func() as
2238 func(void). Also stripped out the parts of mkdef.pl that
2239 added extra typesafe functions: these no longer exist.
2242 *) Reorganisation of the stack code. The macros are now all
2243 collected in safestack.h . Each macro is defined in terms of
2244 a "stack macro" of the form SKM_<name>(type, a, b). The
2245 DEBUG_SAFESTACK is now handled in terms of function casts,
2246 this has the advantage of retaining type safety without the
2247 use of additional functions. If DEBUG_SAFESTACK is not defined
2248 then the non typesafe macros are used instead. Also modified the
2249 mkstack.pl script to handle the new form. Needs testing to see
2250 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
2251 the default if no major problems. Similar behaviour for ASN1_SET_OF
2252 and PKCS12_STACK_OF.
2255 *) When some versions of IIS use the 'NET' form of private key the
2256 key derivation algorithm is different. Normally MD5(password) is
2257 used as a 128 bit RC4 key. In the modified case
2258 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
2259 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
2260 as the old Netscape_RSA functions except they have an additional
2261 'sgckey' parameter which uses the modified algorithm. Also added
2262 an -sgckey command line option to the rsa utility. Thanks to
2263 Adrian Peck <bertie@ncipher.com> for posting details of the modified
2264 algorithm to openssl-dev.
2267 *) The evp_local.h macros were using 'c.##kname' which resulted in
2268 invalid expansion on some systems (SCO 5.0.5 for example).
2269 Corrected to 'c.kname'.
2270 [Phillip Porch <root@theporch.com>]
2272 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
2273 a STACK of email addresses from a certificate or request, these look
2274 in the subject name and the subject alternative name extensions and
2275 omit any duplicate addresses.
2278 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
2279 This makes DSA verification about 2 % faster.
2282 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
2283 (meaning that now 2^5 values will be precomputed, which is only 4 KB
2284 plus overhead for 1024 bit moduli).
2285 This makes exponentiations about 0.5 % faster for 1024 bit
2286 exponents (as measured by "openssl speed rsa2048").
2289 *) Rename memory handling macros to avoid conflicts with other
2291 Malloc => OPENSSL_malloc
2292 Malloc_locked => OPENSSL_malloc_locked
2293 Realloc => OPENSSL_realloc
2294 Free => OPENSSL_free
2297 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
2298 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
2301 *) CygWin32 support.
2302 [John Jarvie <jjarvie@newsguy.com>]
2304 *) The type-safe stack code has been rejigged. It is now only compiled
2305 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
2306 by default all type-specific stack functions are "#define"d back to
2307 standard stack functions. This results in more streamlined output
2308 but retains the type-safety checking possibilities of the original
2312 *) The STACK code has been cleaned up, and certain type declarations
2313 that didn't make a lot of sense have been brought in line. This has
2314 also involved a cleanup of sorts in safestack.h to more correctly
2315 map type-safe stack functions onto their plain stack counterparts.
2316 This work has also resulted in a variety of "const"ifications of
2317 lots of the code, especially "_cmp" operations which should normally
2318 be prototyped with "const" parameters anyway.
2321 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
2322 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
2323 (The PRNG state consists of two parts, the large pool 'state' and 'md',
2324 where all of 'md' is used each time the PRNG is used, but 'state'
2325 is used only indexed by a cyclic counter. As entropy may not be
2326 well distributed from the beginning, 'md' is important as a
2327 chaining variable. However, the output function chains only half
2328 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
2329 all of 'md', and seeding with STATE_SIZE dummy bytes will result
2330 in all of 'state' being rewritten, with the new values depending
2331 on virtually all of 'md'. This overcomes the 80 bit limitation.)
2334 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
2335 the handshake is continued after ssl_verify_cert_chain();
2336 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
2337 can lead to 'unexplainable' connection aborts later.
2338 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
2340 *) Major EVP API cipher revision.
2341 Add hooks for extra EVP features. This allows various cipher
2342 parameters to be set in the EVP interface. Support added for variable
2343 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
2344 setting of RC2 and RC5 parameters.
2346 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
2349 Remove lots of duplicated code from the EVP library. For example *every*
2350 cipher init() function handles the 'iv' in the same way according to the
2351 cipher mode. They also all do nothing if the 'key' parameter is NULL and
2352 for CFB and OFB modes they zero ctx->num.
2354 New functionality allows removal of S/MIME code RC2 hack.
2356 Most of the routines have the same form and so can be declared in terms
2359 By shifting this to the top level EVP_CipherInit() it can be removed from
2360 all individual ciphers. If the cipher wants to handle IVs or keys
2361 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
2364 Change lots of functions like EVP_EncryptUpdate() to now return a
2365 value: although software versions of the algorithms cannot fail
2366 any installed hardware versions can.
2369 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
2370 this option is set, tolerate broken clients that send the negotiated
2371 protocol version number instead of the requested protocol version
2375 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
2376 i.e. non-zero for export ciphersuites, zero otherwise.
2377 Previous versions had this flag inverted, inconsistent with
2378 rsa_tmp_cb (..._TMP_RSA_CB).
2379 [Bodo Moeller; problem reported by Amit Chopra]
2381 *) Add missing DSA library text string. Work around for some IIS
2382 key files with invalid SEQUENCE encoding.
2385 *) Add a document (doc/standards.txt) that list all kinds of standards
2386 and so on that are implemented in OpenSSL.
2389 *) Enhance c_rehash script. Old version would mishandle certificates
2390 with the same subject name hash and wouldn't handle CRLs at all.
2391 Added -fingerprint option to crl utility, to support new c_rehash
2395 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
2398 *) Fix for SSL server purpose checking. Server checking was
2399 rejecting certificates which had extended key usage present
2400 but no ssl client purpose.
2401 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
2403 *) Make PKCS#12 code work with no password. The PKCS#12 spec
2404 is a little unclear about how a blank password is handled.
2405 Since the password in encoded as a BMPString with terminating
2406 double NULL a zero length password would end up as just the
2407 double NULL. However no password at all is different and is
2408 handled differently in the PKCS#12 key generation code. NS
2409 treats a blank password as zero length. MSIE treats it as no
2410 password on export: but it will try both on import. We now do
2411 the same: PKCS12_parse() tries zero length and no password if
2412 the password is set to "" or NULL (NULL is now a valid password:
2413 it wasn't before) as does the pkcs12 application.
2416 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
2417 perror when PEM_read_bio_X509_REQ fails, the error message must
2418 be obtained from the error queue.
2421 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
2422 it in ERR_remove_state if appropriate, and change ERR_get_state
2423 accordingly to avoid race conditions (this is necessary because
2424 thread_hash is no longer constant once set).
2427 *) Bugfix for linux-elf makefile.one.
2430 *) RSA_get_default_method() will now cause a default
2431 RSA_METHOD to be chosen if one doesn't exist already.
2432 Previously this was only set during a call to RSA_new()
2433 or RSA_new_method(NULL) meaning it was possible for
2434 RSA_get_default_method() to return NULL.
2437 *) Added native name translation to the existing DSO code
2438 that will convert (if the flag to do so is set) filenames
2439 that are sufficiently small and have no path information
2440 into a canonical native form. Eg. "blah" converted to
2441 "libblah.so" or "blah.dll" etc.
2444 *) New function ERR_error_string_n(e, buf, len) which is like
2445 ERR_error_string(e, buf), but writes at most 'len' bytes
2446 including the 0 terminator. For ERR_error_string_n, 'buf'
2448 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
2450 *) CONF library reworked to become more general. A new CONF
2451 configuration file reader "class" is implemented as well as a
2452 new functions (NCONF_*, for "New CONF") to handle it. The now
2453 old CONF_* functions are still there, but are reimplemented to
2454 work in terms of the new functions. Also, a set of functions
2455 to handle the internal storage of the configuration data is
2456 provided to make it easier to write new configuration file
2457 reader "classes" (I can definitely see something reading a
2458 configuration file in XML format, for example), called _CONF_*,
2459 or "the configuration storage API"...
2461 The new configuration file reading functions are:
2463 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
2464 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
2466 NCONF_default, NCONF_WIN32
2468 NCONF_dump_fp, NCONF_dump_bio
2470 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
2471 NCONF_new creates a new CONF object. This works in the same way
2472 as other interfaces in OpenSSL, like the BIO interface.
2473 NCONF_dump_* dump the internal storage of the configuration file,
2474 which is useful for debugging. All other functions take the same
2475 arguments as the old CONF_* functions wth the exception of the
2476 first that must be a `CONF *' instead of a `LHASH *'.
2478 To make it easer to use the new classes with the old CONF_* functions,
2479 the function CONF_set_default_method is provided.
2482 *) Add '-tls1' option to 'openssl ciphers', which was already
2483 mentioned in the documentation but had not been implemented.
2484 (This option is not yet really useful because even the additional
2485 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
2488 *) Initial DSO code added into libcrypto for letting OpenSSL (and
2489 OpenSSL-based applications) load shared libraries and bind to
2490 them in a portable way.
2491 [Geoff Thorpe, with contributions from Richard Levitte]
2493 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
2495 *) Make sure _lrotl and _lrotr are only used with MSVC.
2497 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
2498 (the default implementation of RAND_status).
2500 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
2501 to '-clrext' (= clear extensions), as intended and documented.
2502 [Bodo Moeller; inconsistency pointed out by Michael Attili
2505 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
2506 was larger than the MD block size.
2507 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
2509 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
2510 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
2511 using the passed key: if the passed key was a private key the result
2512 of X509_print(), for example, would be to print out all the private key
2516 *) des_quad_cksum() byte order bug fix.
2517 [Ulf Möller, using the problem description in krb4-0.9.7, where
2518 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
2520 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
2522 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
2524 *) For easily testing in shell scripts whether some command
2525 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
2526 returns with exit code 0 iff no command of the given name is available.
2527 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
2528 the output goes to stdout and nothing is printed to stderr.
2529 Additional arguments are always ignored.
2531 Since for each cipher there is a command of the same name,
2532 the 'no-cipher' compilation switches can be tested this way.
2534 ('openssl no-XXX' is not able to detect pseudo-commands such
2535 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
2538 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
2541 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
2542 is set; it will be thrown away anyway because each handshake creates
2544 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
2545 to parameters -- in previous versions (since OpenSSL 0.9.3) the
2546 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
2547 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
2550 *) New s_client option -ign_eof: EOF at stdin is ignored, and
2551 'Q' and 'R' lose their special meanings (quit/renegotiate).
2552 This is part of what -quiet does; unlike -quiet, -ign_eof
2553 does not suppress any output.
2556 *) Add compatibility options to the purpose and trust code. The
2557 purpose X509_PURPOSE_ANY is "any purpose" which automatically
2558 accepts a certificate or CA, this was the previous behaviour,
2559 with all the associated security issues.
2561 X509_TRUST_COMPAT is the old trust behaviour: only and
2562 automatically trust self signed roots in certificate store. A
2563 new trust setting X509_TRUST_DEFAULT is used to specify that
2564 a purpose has no associated trust setting and it should instead
2565 use the value in the default purpose.
2568 *) Fix the PKCS#8 DSA private key code so it decodes keys again
2569 and fix a memory leak.
2572 *) In util/mkerr.pl (which implements 'make errors'), preserve
2573 reason strings from the previous version of the .c file, as
2574 the default to have only downcase letters (and digits) in
2575 automatically generated reasons codes is not always appropriate.
2578 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
2579 using strerror. Previously, ERR_reason_error_string() returned
2580 library names as reason strings for SYSerr; but SYSerr is a special
2581 case where small numbers are errno values, not library numbers.
2584 *) Add '-dsaparam' option to 'openssl dhparam' application. This
2585 converts DSA parameters into DH parameters. (When creating parameters,
2586 DSA_generate_parameters is used.)
2589 *) Include 'length' (recommended exponent length) in C code generated
2590 by 'openssl dhparam -C'.
2593 *) The second argument to set_label in perlasm was already being used
2594 so couldn't be used as a "file scope" flag. Moved to third argument
2598 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
2599 instead of RAND_bytes for encryption IVs and salts.
2602 *) Include RAND_status() into RAND_METHOD instead of implementing
2603 it only for md_rand.c Otherwise replacing the PRNG by calling
2604 RAND_set_rand_method would be impossible.
2607 *) Don't let DSA_generate_key() enter an infinite loop if the random
2608 number generation fails.
2611 *) New 'rand' application for creating pseudo-random output.
2614 *) Added configuration support for Linux/IA64
2615 [Rolf Haberrecker <rolf@suse.de>]
2617 *) Assembler module support for Mingw32.
2620 *) Shared library support for HPUX (in shlib/).
2621 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
2623 *) Shared library support for Solaris gcc.
2624 [Lutz Behnke <behnke@trustcenter.de>]
2626 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
2628 *) PKCS7_encrypt() was adding text MIME headers twice because they
2629 were added manually and by SMIME_crlf_copy().
2632 *) In bntest.c don't call BN_rand with zero bits argument.
2633 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
2635 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
2636 case was implemented. This caused BN_div_recp() to fail occasionally.
2639 *) Add an optional second argument to the set_label() in the perl
2640 assembly language builder. If this argument exists and is set
2641 to 1 it signals that the assembler should use a symbol whose
2642 scope is the entire file, not just the current function. This
2643 is needed with MASM which uses the format label:: for this scope.
2644 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
2646 *) Change the ASN1 types so they are typedefs by default. Before
2647 almost all types were #define'd to ASN1_STRING which was causing
2648 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
2652 *) Change names of new functions to the new get1/get0 naming
2653 convention: After 'get1', the caller owns a reference count
2654 and has to call ..._free; 'get0' returns a pointer to some
2655 data structure without incrementing reference counters.
2656 (Some of the existing 'get' functions increment a reference
2657 counter, some don't.)
2658 Similarly, 'set1' and 'add1' functions increase reference
2659 counters or duplicate objects.
2662 *) Allow for the possibility of temp RSA key generation failure:
2663 the code used to assume it always worked and crashed on failure.
2666 *) Fix potential buffer overrun problem in BIO_printf().
2667 [Ulf Möller, using public domain code by Patrick Powell; problem
2668 pointed out by David Sacerdote <das33@cornell.edu>]
2670 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
2671 RAND_egd() and RAND_status(). In the command line application,
2672 the EGD socket can be specified like a seed file using RANDFILE
2676 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
2677 Some CAs (e.g. Verisign) distribute certificates in this form.
2680 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
2681 list to exclude them. This means that no special compilation option
2682 is needed to use anonymous DH: it just needs to be included in the
2686 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
2687 EVP_MD_type. The old functionality is available in a new macro called
2688 EVP_MD_md(). Change code that uses it and update docs.
2691 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
2692 where the 'void *' argument is replaced by a function pointer argument.
2693 Previously 'void *' was abused to point to functions, which works on
2694 many platforms, but is not correct. As these functions are usually
2695 called by macros defined in OpenSSL header files, most source code
2696 should work without changes.
2699 *) <openssl/opensslconf.h> (which is created by Configure) now contains
2700 sections with information on -D... compiler switches used for
2701 compiling the library so that applications can see them. To enable
2702 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
2703 must be defined. E.g.,
2704 #define OPENSSL_ALGORITHM_DEFINES
2705 #include <openssl/opensslconf.h>
2706 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
2707 [Richard Levitte, Ulf and Bodo Möller]
2709 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
2713 *) Change the 'other' type in certificate aux info to a STACK_OF
2714 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
2715 the required ASN1 format: arbitrary types determined by an OID.
2718 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
2719 argument to 'req'. This is not because the function is newer or
2720 better than others it just uses the work 'NEW' in the certificate
2721 request header lines. Some software needs this.
2724 *) Reorganise password command line arguments: now passwords can be
2725 obtained from various sources. Delete the PEM_cb function and make
2726 it the default behaviour: i.e. if the callback is NULL and the
2727 usrdata argument is not NULL interpret it as a null terminated pass
2728 phrase. If usrdata and the callback are NULL then the pass phrase
2729 is prompted for as usual.
2732 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
2733 the support is automatically enabled. The resulting binaries will
2734 autodetect the card and use it if present.
2735 [Ben Laurie and Compaq Inc.]
2737 *) Work around for Netscape hang bug. This sends certificate request
2738 and server done in one record. Since this is perfectly legal in the
2739 SSL/TLS protocol it isn't a "bug" option and is on by default. See
2740 the bugs/SSLv3 entry for more info.
2743 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
2746 *) Add -rand argument to smime and pkcs12 applications and read/write
2750 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
2753 *) Add command line password options to the remaining applications.
2756 *) Bug fix for BN_div_recp() for numerators with an even number of
2760 *) More tests in bntest.c, and changed test_bn output.
2763 *) ./config recognizes MacOS X now.
2766 *) Bug fix for BN_div() when the first words of num and divsor are
2767 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
2770 *) Add support for various broken PKCS#8 formats, and command line
2771 options to produce them.
2774 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
2775 get temporary BIGNUMs from a BN_CTX.
2778 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
2782 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
2783 include a #define from the old name to the new. The original intent
2784 was that statically linked binaries could for example just call
2785 SSLeay_add_all_ciphers() to just add ciphers to the table and not
2786 link with digests. This never worked becayse SSLeay_add_all_digests()
2787 and SSLeay_add_all_ciphers() were in the same source file so calling
2788 one would link with the other. They are now in separate source files.
2791 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
2794 *) Use a less unusual form of the Miller-Rabin primality test (it used
2795 a binary algorithm for exponentiation integrated into the Miller-Rabin
2796 loop, our standard modexp algorithms are faster).
2799 *) Support for the EBCDIC character set completed.
2800 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
2802 *) Source code cleanups: use const where appropriate, eliminate casts,
2803 use void * instead of char * in lhash.
2806 *) Bugfix: ssl3_send_server_key_exchange was not restartable
2807 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
2808 this the server could overwrite ephemeral keys that the client
2812 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
2813 using 50 iterations of the Rabin-Miller test.
2815 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
2816 iterations of the Rabin-Miller test as required by the appendix
2817 to FIPS PUB 186[-1]) instead of DSA_is_prime.
2818 As BN_is_prime_fasttest includes trial division, DSA parameter
2819 generation becomes much faster.
2821 This implies a change for the callback functions in DSA_is_prime
2822 and DSA_generate_parameters: The callback function is called once
2823 for each positive witness in the Rabin-Miller test, not just
2824 occasionally in the inner loop; and the parameters to the
2825 callback function now provide an iteration count for the outer
2826 loop rather than for the current invocation of the inner loop.
2827 DSA_generate_parameters additionally can call the callback
2828 function with an 'iteration count' of -1, meaning that a
2829 candidate has passed the trial division test (when q is generated
2830 from an application-provided seed, trial division is skipped).
2833 *) New function BN_is_prime_fasttest that optionally does trial
2834 division before starting the Rabin-Miller test and has
2835 an additional BN_CTX * argument (whereas BN_is_prime always
2836 has to allocate at least one BN_CTX).
2837 'callback(1, -1, cb_arg)' is called when a number has passed the
2838 trial division stage.
2841 *) Fix for bug in CRL encoding. The validity dates weren't being handled
2845 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
2848 *) New function BN_pseudo_rand().
2851 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
2852 bignum version of BN_from_montgomery() with the working code from
2853 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
2857 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
2858 made it impossible to use the same SSL_SESSION data structure in
2859 SSL2 clients in multiple threads.
2862 *) The return value of RAND_load_file() no longer counts bytes obtained
2863 by stat(). RAND_load_file(..., -1) is new and uses the complete file
2864 to seed the PRNG (previously an explicit byte count was required).
2865 [Ulf Möller, Bodo Möller]
2867 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
2868 used (char *) instead of (void *) and had casts all over the place.
2871 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
2874 *) Retain source code compatibility for BN_prime_checks macro:
2875 BN_is_prime(..., BN_prime_checks, ...) now uses
2876 BN_prime_checks_for_size to determine the appropriate number of
2877 Rabin-Miller iterations.
2880 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
2881 DH_CHECK_P_NOT_SAFE_PRIME.
2882 (Check if this is true? OpenPGP calls them "strong".)
2885 *) Merge the functionality of "dh" and "gendh" programs into a new program
2886 "dhparam". The old programs are retained for now but will handle DH keys
2887 (instead of parameters) in future.
2890 *) Make the ciphers, s_server and s_client programs check the return values
2891 when a new cipher list is set.
2894 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
2895 ciphers. Before when the 56bit ciphers were enabled the sorting was
2898 The syntax for the cipher sorting has been extended to support sorting by
2899 cipher-strength (using the strength_bits hard coded in the tables).
2900 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
2902 Fix a bug in the cipher-command parser: when supplying a cipher command
2903 string with an "undefined" symbol (neither command nor alphanumeric
2904 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
2905 an error is flagged.
2907 Due to the strength-sorting extension, the code of the
2908 ssl_create_cipher_list() function was completely rearranged. I hope that
2909 the readability was also increased :-)
2910 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
2912 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
2913 for the first serial number and places 2 in the serial number file. This
2914 avoids problems when the root CA is created with serial number zero and
2915 the first user certificate has the same issuer name and serial number
2919 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
2920 the new code. Add documentation for this stuff.
2923 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
2924 X509_*() to X509at_*() on the grounds that they don't handle X509
2925 structures and behave in an analagous way to the X509v3 functions:
2926 they shouldn't be called directly but wrapper functions should be used
2929 So we also now have some wrapper functions that call the X509at functions
2930 when passed certificate requests. (TO DO: similar things can be done with
2931 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
2932 things. Some of these need some d2i or i2d and print functionality
2933 because they handle more complex structures.)
2936 *) Add missing #ifndefs that caused missing symbols when building libssl
2937 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
2938 NO_RSA in ssl/s2*.c.
2939 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
2941 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
2942 has a return value which indicates the quality of the random data
2943 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
2944 error queue. New function RAND_pseudo_bytes() generates output that is
2945 guaranteed to be unique but not unpredictable. RAND_add is like
2946 RAND_seed, but takes an extra argument for an entropy estimate
2947 (RAND_seed always assumes full entropy).
2950 *) Do more iterations of Rabin-Miller probable prime test (specifically,
2951 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
2952 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
2953 in crypto/bn/bn_prime.c for the complete table). This guarantees a
2954 false-positive rate of at most 2^-80 for random input.
2957 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
2960 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
2961 in the 0.9.5 release), this returns the chain
2962 from an X509_CTX structure with a dup of the stack and all
2963 the X509 reference counts upped: so the stack will exist
2964 after X509_CTX_cleanup() has been called. Modify pkcs12.c
2967 Also make SSL_SESSION_print() print out the verify return
2971 *) Add manpage for the pkcs12 command. Also change the default
2972 behaviour so MAC iteration counts are used unless the new
2973 -nomaciter option is used. This improves file security and
2974 only older versions of MSIE (4.0 for example) need it.
2977 *) Honor the no-xxx Configure options when creating .DEF files.
2980 *) Add PKCS#10 attributes to field table: challengePassword,
2981 unstructuredName and unstructuredAddress. These are taken from
2982 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
2983 international characters are used.
2985 More changes to X509_ATTRIBUTE code: allow the setting of types
2986 based on strings. Remove the 'loc' parameter when adding
2987 attributes because these will be a SET OF encoding which is sorted
2991 *) Initial changes to the 'req' utility to allow request generation
2992 automation. This will allow an application to just generate a template
2993 file containing all the field values and have req construct the
2996 Initial support for X509_ATTRIBUTE handling. Stacks of these are
2997 used all over the place including certificate requests and PKCS#7
2998 structures. They are currently handled manually where necessary with
2999 some primitive wrappers for PKCS#7. The new functions behave in a
3000 manner analogous to the X509 extension functions: they allow
3001 attributes to be looked up by NID and added.
3003 Later something similar to the X509V3 code would be desirable to
3004 automatically handle the encoding, decoding and printing of the
3005 more complex types. The string types like challengePassword can
3006 be handled by the string table functions.
3008 Also modified the multi byte string table handling. Now there is
3009 a 'global mask' which masks out certain types. The table itself
3010 can use the flag STABLE_NO_MASK to ignore the mask setting: this
3011 is useful when for example there is only one permissible type
3012 (as in countryName) and using the mask might result in no valid
3016 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
3017 SSL_get_peer_finished to allow applications to obtain the latest
3018 Finished messages sent to the peer or expected from the peer,
3019 respectively. (SSL_get_peer_finished is usually the Finished message
3020 actually received from the peer, otherwise the protocol will be aborted.)
3022 As the Finished message are message digests of the complete handshake
3023 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
3024 be used for external authentication procedures when the authentication
3025 provided by SSL/TLS is not desired or is not enough.
3028 *) Enhanced support for Alpha Linux is added. Now ./config checks if
3029 the host supports BWX extension and if Compaq C is present on the
3030 $PATH. Just exploiting of the BWX extension results in 20-30%
3031 performance kick for some algorithms, e.g. DES and RC4 to mention
3032 a couple. Compaq C in turn generates ~20% faster code for MD5 and
3036 *) Add support for MS "fast SGC". This is arguably a violation of the
3037 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
3038 weak crypto and after checking the certificate is SGC a second one
3039 with strong crypto. MS SGC stops the first handshake after receiving
3040 the server certificate message and sends a second client hello. Since
3041 a server will typically do all the time consuming operations before
3042 expecting any further messages from the client (server key exchange
3043 is the most expensive) there is little difference between the two.
3045 To get OpenSSL to support MS SGC we have to permit a second client
3046 hello message after we have sent server done. In addition we have to
3047 reset the MAC if we do get this second client hello.
3050 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
3051 if a DER encoded private key is RSA or DSA traditional format. Changed
3052 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
3053 format DER encoded private key. Newer code should use PKCS#8 format which
3054 has the key type encoded in the ASN1 structure. Added DER private key
3055 support to pkcs8 application.
3058 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
3059 ciphersuites has been selected (as required by the SSL 3/TLS 1
3060 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
3061 is set, we interpret this as a request to violate the specification
3062 (the worst that can happen is a handshake failure, and 'correct'
3063 behaviour would result in a handshake failure anyway).
3066 *) In SSL_CTX_add_session, take into account that there might be multiple
3067 SSL_SESSION structures with the same session ID (e.g. when two threads
3068 concurrently obtain them from an external cache).
3069 The internal cache can handle only one SSL_SESSION with a given ID,
3070 so if there's a conflict, we now throw out the old one to achieve
3074 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
3075 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
3076 some routines that use cipher OIDs: some ciphers do not have OIDs
3077 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
3081 *) Simplify the trust setting structure and code. Now we just have
3082 two sequences of OIDs for trusted and rejected settings. These will
3083 typically have values the same as the extended key usage extension
3084 and any application specific purposes.
3086 The trust checking code now has a default behaviour: it will just
3087 check for an object with the same NID as the passed id. Functions can
3088 be provided to override either the default behaviour or the behaviour
3089 for a given id. SSL client, server and email already have functions
3090 in place for compatibility: they check the NID and also return "trusted"
3091 if the certificate is self signed.
3094 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
3095 traditional format into an EVP_PKEY structure.
3098 *) Add a password callback function PEM_cb() which either prompts for
3099 a password if usr_data is NULL or otherwise assumes it is a null
3100 terminated password. Allow passwords to be passed on command line
3101 environment or config files in a few more utilities.
3104 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
3105 keys. Add some short names for PKCS#8 PBE algorithms and allow them
3106 to be specified on the command line for the pkcs8 and pkcs12 utilities.
3107 Update documentation.
3110 *) Support for ASN1 "NULL" type. This could be handled before by using
3111 ASN1_TYPE but there wasn't any function that would try to read a NULL
3112 and produce an error if it couldn't. For compatibility we also have
3113 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
3114 don't allocate anything because they don't need to.
3117 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
3119 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
3121 *) Rebuild of the memory allocation routines used by OpenSSL code and
3122 possibly others as well. The purpose is to make an interface that
3123 provide hooks so anyone can build a separate set of allocation and
3124 deallocation routines to be used by OpenSSL, for example memory
3125 pool implementations, or something else, which was previously hard
3126 since Malloc(), Realloc() and Free() were defined as macros having
3127 the values malloc, realloc and free, respectively (except for Win32
3128 compilations). The same is provided for memory debugging code.
3129 OpenSSL already comes with functionality to find memory leaks, but
3130 this gives people a chance to debug other memory problems.
3132 With these changes, a new set of functions and macros have appeared:
3134 CRYPTO_set_mem_debug_functions() [F]
3135 CRYPTO_get_mem_debug_functions() [F]
3136 CRYPTO_dbg_set_options() [F]
3137 CRYPTO_dbg_get_options() [F]
3138 CRYPTO_malloc_debug_init() [M]
3140 The memory debug functions are NULL by default, unless the library
3141 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
3142 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
3143 gives the standard debugging functions that come with OpenSSL) or
3144 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
3145 provided by the library user) must be used. When the standard
3146 debugging functions are used, CRYPTO_dbg_set_options can be used to
3147 request additional information:
3148 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
3149 the CRYPTO_MDEBUG_xxx macro when compiling the library.
3151 Also, things like CRYPTO_set_mem_functions will always give the
3152 expected result (the new set of functions is used for allocation
3153 and deallocation) at all times, regardless of platform and compiler
3156 To finish it up, some functions that were never use in any other
3157 way than through macros have a new API and new semantic:
3160 CRYPTO_dbg_realloc()
3163 All macros of value have retained their old syntax.
3164 [Richard Levitte and Bodo Moeller]
3166 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
3167 ordering of SMIMECapabilities wasn't in "strength order" and there
3168 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
3172 *) Some ASN1 types with illegal zero length encoding (INTEGER,
3173 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
3174 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
3176 *) Merge in my S/MIME library for OpenSSL. This provides a simple
3177 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
3178 functionality to handle multipart/signed properly) and a utility
3179 called 'smime' to call all this stuff. This is based on code I
3180 originally wrote for Celo who have kindly allowed it to be
3181 included in OpenSSL.
3184 *) Add variants des_set_key_checked and des_set_key_unchecked of
3185 des_set_key (aka des_key_sched). Global variable des_check_key
3186 decides which of these is called by des_set_key; this way
3187 des_check_key behaves as it always did, but applications and
3188 the library itself, which was buggy for des_check_key == 1,
3189 have a cleaner way to pick the version they need.
3192 *) New function PKCS12_newpass() which changes the password of a
3196 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
3197 dynamic mix. In both cases the ids can be used as an index into the
3198 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
3199 functions so they accept a list of the field values and the
3200 application doesn't need to directly manipulate the X509_TRUST
3204 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
3208 *) Modify the way the V3 extension code looks up extensions. This now
3209 works in a similar way to the object code: we have some "standard"
3210 extensions in a static table which is searched with OBJ_bsearch()
3211 and the application can add dynamic ones if needed. The file
3212 crypto/x509v3/ext_dat.h now has the info: this file needs to be
3213 updated whenever a new extension is added to the core code and kept
3214 in ext_nid order. There is a simple program 'tabtest.c' which checks
3215 this. New extensions are not added too often so this file can readily
3216 be maintained manually.
3218 There are two big advantages in doing things this way. The extensions
3219 can be looked up immediately and no longer need to be "added" using
3220 X509V3_add_standard_extensions(): this function now does nothing.
3221 [Side note: I get *lots* of email saying the extension code doesn't
3222 work because people forget to call this function]
3223 Also no dynamic allocation is done unless new extensions are added:
3224 so if we don't add custom extensions there is no need to call
3225 X509V3_EXT_cleanup().
3228 *) Modify enc utility's salting as follows: make salting the default. Add a
3229 magic header, so unsalted files fail gracefully instead of just decrypting
3230 to garbage. This is because not salting is a big security hole, so people
3231 should be discouraged from doing it.
3234 *) Fixes and enhancements to the 'x509' utility. It allowed a message
3235 digest to be passed on the command line but it only used this
3236 parameter when signing a certificate. Modified so all relevant
3237 operations are affected by the digest parameter including the
3238 -fingerprint and -x509toreq options. Also -x509toreq choked if a
3239 DSA key was used because it didn't fix the digest.
3242 *) Initial certificate chain verify code. Currently tests the untrusted
3243 certificates for consistency with the verify purpose (which is set
3244 when the X509_STORE_CTX structure is set up) and checks the pathlength.
3246 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
3247 this is because it will reject chains with invalid extensions whereas
3248 every previous version of OpenSSL and SSLeay made no checks at all.
3250 Trust code: checks the root CA for the relevant trust settings. Trust
3251 settings have an initial value consistent with the verify purpose: e.g.
3252 if the verify purpose is for SSL client use it expects the CA to be
3253 trusted for SSL client use. However the default value can be changed to
3254 permit custom trust settings: one example of this would be to only trust
3255 certificates from a specific "secure" set of CAs.
3257 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
3258 which should be used for version portability: especially since the
3259 verify structure is likely to change more often now.
3261 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
3262 to set them. If not set then assume SSL clients will verify SSL servers
3265 Two new options to the verify program: -untrusted allows a set of
3266 untrusted certificates to be passed in and -purpose which sets the
3267 intended purpose of the certificate. If a purpose is set then the
3268 new chain verify code is used to check extension consistency.
3271 *) Support for the authority information access extension.
3274 *) Modify RSA and DSA PEM read routines to transparently handle
3275 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
3276 public keys in a format compatible with certificate
3277 SubjectPublicKeyInfo structures. Unfortunately there were already
3278 functions called *_PublicKey_* which used various odd formats so
3279 these are retained for compatibility: however the DSA variants were
3280 never in a public release so they have been deleted. Changed dsa/rsa
3281 utilities to handle the new format: note no releases ever handled public
3282 keys so we should be OK.
3284 The primary motivation for this change is to avoid the same fiasco
3285 that dogs private keys: there are several incompatible private key
3286 formats some of which are standard and some OpenSSL specific and
3287 require various evil hacks to allow partial transparent handling and
3288 even then it doesn't work with DER formats. Given the option anything
3289 other than PKCS#8 should be dumped: but the other formats have to
3290 stay in the name of compatibility.
3292 With public keys and the benefit of hindsight one standard format
3293 is used which works with EVP_PKEY, RSA or DSA structures: though
3294 it clearly returns an error if you try to read the wrong kind of key.
3296 Added a -pubkey option to the 'x509' utility to output the public key.
3297 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
3298 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
3299 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
3300 that do the same as the EVP_PKEY_assign_*() except they up the
3301 reference count of the added key (they don't "swallow" the
3305 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
3306 CRLs would fail if the file contained no certificates or no CRLs:
3307 added a new function to read in both types and return the number
3308 read: this means that if none are read it will be an error. The
3309 DER versions of the certificate and CRL reader would always fail
3310 because it isn't possible to mix certificates and CRLs in DER format
3311 without choking one or the other routine. Changed this to just read
3312 a certificate: this is the best we can do. Also modified the code
3313 in apps/verify.c to take notice of return codes: it was previously
3314 attempting to read in certificates from NULL pointers and ignoring
3315 any errors: this is one reason why the cert and CRL reader seemed
3316 to work. It doesn't check return codes from the default certificate
3317 routines: these may well fail if the certificates aren't installed.
3320 *) Code to support otherName option in GeneralName.
3323 *) First update to verify code. Change the verify utility
3324 so it warns if it is passed a self signed certificate:
3325 for consistency with the normal behaviour. X509_verify
3326 has been modified to it will now verify a self signed
3327 certificate if *exactly* the same certificate appears
3328 in the store: it was previously impossible to trust a
3329 single self signed certificate. This means that:
3330 openssl verify ss.pem
3331 now gives a warning about a self signed certificate but
3332 openssl verify -CAfile ss.pem ss.pem
3336 *) For servers, store verify_result in SSL_SESSION data structure
3337 (and add it to external session representation).
3338 This is needed when client certificate verifications fails,
3339 but an application-provided verification callback (set by
3340 SSL_CTX_set_cert_verify_callback) allows accepting the session
3341 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
3342 but returns 1): When the session is reused, we have to set
3343 ssl->verify_result to the appropriate error code to avoid
3345 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
3347 *) Fix a bug in the new PKCS#7 code: it didn't consider the
3348 case in PKCS7_dataInit() where the signed PKCS7 structure
3349 didn't contain any existing data because it was being created.
3350 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
3352 *) Add a salt to the key derivation routines in enc.c. This
3353 forms the first 8 bytes of the encrypted file. Also add a
3354 -S option to allow a salt to be input on the command line.
3357 *) New function X509_cmp(). Oddly enough there wasn't a function
3358 to compare two certificates. We do this by working out the SHA1
3359 hash and comparing that. X509_cmp() will be needed by the trust
3363 *) SSL_get1_session() is like SSL_get_session(), but increments
3364 the reference count in the SSL_SESSION returned.
3365 [Geoff Thorpe <geoff@eu.c2.net>]
3367 *) Fix for 'req': it was adding a null to request attributes.
3368 Also change the X509_LOOKUP and X509_INFO code to handle
3369 certificate auxiliary information.
3372 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
3376 *) Add the possibility to add extra information to the memory leak
3377 detecting output, to form tracebacks, showing from where each
3378 allocation was originated: CRYPTO_push_info("constant string") adds
3379 the string plus current file name and line number to a per-thread
3380 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
3381 is like calling CYRPTO_pop_info() until the stack is empty.
3382 Also updated memory leak detection code to be multi-thread-safe.
3385 *) Add options -text and -noout to pkcs7 utility and delete the
3386 encryption options which never did anything. Update docs.
3389 *) Add options to some of the utilities to allow the pass phrase
3390 to be included on either the command line (not recommended on
3391 OSes like Unix) or read from the environment. Update the
3392 manpages and fix a few bugs.
3395 *) Add a few manpages for some of the openssl commands.
3398 *) Fix the -revoke option in ca. It was freeing up memory twice,
3399 leaking and not finding already revoked certificates.
3402 *) Extensive changes to support certificate auxiliary information.
3403 This involves the use of X509_CERT_AUX structure and X509_AUX
3404 functions. An X509_AUX function such as PEM_read_X509_AUX()
3405 can still read in a certificate file in the usual way but it
3406 will also read in any additional "auxiliary information". By
3407 doing things this way a fair degree of compatibility can be
3408 retained: existing certificates can have this information added
3409 using the new 'x509' options.
3411 Current auxiliary information includes an "alias" and some trust
3412 settings. The trust settings will ultimately be used in enhanced
3413 certificate chain verification routines: currently a certificate
3414 can only be trusted if it is self signed and then it is trusted
3418 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
3419 The problem was that one of the replacement routines had not been working
3420 since SSLeay releases. For now the offending routine has been replaced
3421 with non-optimised assembler. Even so, this now gives around 95%
3422 performance improvement for 1024 bit RSA signs.
3425 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
3426 handling. Most clients have the effective key size in bits equal to
3427 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
3428 A few however don't do this and instead use the size of the decrypted key
3429 to determine the RC2 key length and the AlgorithmIdentifier to determine
3430 the effective key length. In this case the effective key length can still
3431 be 40 bits but the key length can be 168 bits for example. This is fixed
3432 by manually forcing an RC2 key into the EVP_PKEY structure because the
3433 EVP code can't currently handle unusual RC2 key sizes: it always assumes
3434 the key length and effective key length are equal.
3437 *) Add a bunch of functions that should simplify the creation of
3438 X509_NAME structures. Now you should be able to do:
3439 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
3440 and have it automatically work out the correct field type and fill in
3441 the structures. The more adventurous can try:
3442 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
3443 and it will (hopefully) work out the correct multibyte encoding.
3446 *) Change the 'req' utility to use the new field handling and multibyte
3447 copy routines. Before the DN field creation was handled in an ad hoc
3448 way in req, ca, and x509 which was rather broken and didn't support
3449 BMPStrings or UTF8Strings. Since some software doesn't implement
3450 BMPStrings or UTF8Strings yet, they can be enabled using the config file
3451 using the dirstring_type option. See the new comment in the default
3452 openssl.cnf for more info.
3455 *) Make crypto/rand/md_rand.c more robust:
3456 - Assure unique random numbers after fork().
3457 - Make sure that concurrent threads access the global counter and
3458 md serializably so that we never lose entropy in them
3459 or use exactly the same state in multiple threads.
3460 Access to the large state is not always serializable because
3461 the additional locking could be a performance killer, and
3462 md should be large enough anyway.
3465 *) New file apps/app_rand.c with commonly needed functionality
3466 for handling the random seed file.
3468 Use the random seed file in some applications that previously did not:
3470 dsaparam -genkey (which also ignored its '-rand' option),
3473 x509 (when signing).
3474 Except on systems with /dev/urandom, it is crucial to have a random
3475 seed file at least for key creation, DSA signing, and for DH exchanges;
3476 for RSA signatures we could do without one.
3478 gendh and gendsa (unlike genrsa) used to read only the first byte
3479 of each file listed in the '-rand' option. The function as previously
3480 found in genrsa is now in app_rand.c and is used by all programs
3481 that support '-rand'.
3484 *) In RAND_write_file, use mode 0600 for creating files;
3485 don't just chmod when it may be too late.
3488 *) Report an error from X509_STORE_load_locations
3489 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
3492 *) New function ASN1_mbstring_copy() this copies a string in either
3493 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
3494 into an ASN1_STRING type. A mask of permissible types is passed
3495 and it chooses the "minimal" type to use or an error if not type
3499 *) Add function equivalents to the various macros in asn1.h. The old
3500 macros are retained with an M_ prefix. Code inside the library can
3501 use the M_ macros. External code (including the openssl utility)
3502 should *NOT* in order to be "shared library friendly".
3505 *) Add various functions that can check a certificate's extensions
3506 to see if it usable for various purposes such as SSL client,
3507 server or S/MIME and CAs of these types. This is currently
3508 VERY EXPERIMENTAL but will ultimately be used for certificate chain
3509 verification. Also added a -purpose flag to x509 utility to
3510 print out all the purposes.
3513 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
3517 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
3518 for, obtain and decode and extension and obtain its critical flag.
3519 This allows all the necessary extension code to be handled in a
3520 single function call.
3523 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
3524 platforms. See crypto/rc4/rc4_enc.c for further details.
3527 *) New -noout option to asn1parse. This causes no output to be produced
3528 its main use is when combined with -strparse and -out to extract data
3529 from a file (which may not be in ASN.1 format).
3532 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
3533 when producing the local key id.
3534 [Richard Levitte <levitte@stacken.kth.se>]
3536 *) New option -dhparam in s_server. This allows a DH parameter file to be
3537 stated explicitly. If it is not stated then it tries the first server
3538 certificate file. The previous behaviour hard coded the filename
3542 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
3543 a public key to be input or output. For example:
3544 openssl rsa -in key.pem -pubout -out pubkey.pem
3545 Also added necessary DSA public key functions to handle this.
3548 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
3549 in the message. This was handled by allowing
3550 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
3551 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
3553 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
3554 to the end of the strings whereas this didn't. This would cause problems
3555 if strings read with d2i_ASN1_bytes() were later modified.
3556 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
3558 *) Fix for base64 decode bug. When a base64 bio reads only one line of
3559 data and it contains EOF it will end up returning an error. This is
3560 caused by input 46 bytes long. The cause is due to the way base64
3561 BIOs find the start of base64 encoded data. They do this by trying a
3562 trial decode on each line until they find one that works. When they
3563 do a flag is set and it starts again knowing it can pass all the
3564 data directly through the decoder. Unfortunately it doesn
projects / openssl.git / blob