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 developped in parallel, based on OpenSSL 0.9.6.
10 Change log entries are tagged as follows:
11 -) applies to 0.9.6a/0.9.6b/0.9.6c only
12 *) applies to 0.9.6a/0.9.6b/0.9.6c and 0.9.7
13 +) applies to 0.9.7 only
15 +) Add a "destroy" handler to ENGINEs that allows structural cleanup to
16 be done prior to destruction. Use this to unload error strings from
17 ENGINEs that load their own error strings. NB: This adds two new API
18 functions to "get" and "set" this destroy handler in an ENGINE.
21 +) Alter all existing ENGINE implementations (except "openssl" and
22 "openbsd") to dynamically instantiate their own error strings. This
23 makes them more flexible to be built both as statically-linked ENGINEs
24 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
25 Also, add stub code to each that makes building them as self-contained
26 shared-libraries easier (see README.ENGINE).
29 +) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
30 implementations into applications that are completely implemented in
31 self-contained shared-libraries. The "dynamic" ENGINE exposes control
32 commands that can be used to configure what shared-library to load and
33 to control aspects of the way it is handled. Also, made an update to
34 the README.ENGINE file that brings its information up-to-date and
35 provides some information and instructions on the "dynamic" ENGINE
36 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
40 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
42 *) Make it possible to unload ranges of ERR strings with a new
43 "ERR_unload_strings" function.
46 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
47 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
48 followed by modular reduction.
49 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
51 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
52 requivalent based on BN_pseudo_rand() instead of BN_rand().
55 +) Add a copy() function to EVP_MD.
58 +) Make EVP_MD routines take a context pointer instead of just the
62 +) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
63 that the digest can only process a single chunk of data
64 (typically because it is provided by a piece of
65 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
66 is only going to provide a single chunk of data, and hence the
67 framework needn't accumulate the data for oneshot drivers.
70 +) As with "ERR", make it possible to replace the underlying "ex_data"
71 functions. This change also alters the storage and management of global
72 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
73 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
74 index counters. The API functions that use this state have been changed
75 to take a "class_index" rather than pointers to the class's local STACK
76 and counter, and there is now an API function to dynamically create new
77 classes. This centralisation allows us to (a) plug a lot of the
78 thread-safety problems that existed, and (b) makes it possible to clean
79 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
80 such data would previously have always leaked in application code and
81 workarounds were in place to make the memory debugging turn a blind eye
82 to it. Application code that doesn't use this new function will still
83 leak as before, but their memory debugging output will announce it now
84 rather than letting it slide.
86 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
87 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
88 has a return value to indicate success or failure.
91 +) Make it possible to replace the underlying "ERR" functions such that the
92 global state (2 LHASH tables and 2 locks) is only used by the "default"
93 implementation. This change also adds two functions to "get" and "set"
94 the implementation prior to it being automatically set the first time
95 any other ERR function takes place. Ie. an application can call "get",
96 pass the return value to a module it has just loaded, and that module
97 can call its own "set" function using that value. This means the
98 module's "ERR" operations will use (and modify) the error state in the
99 application and not in its own statically linked copy of OpenSSL code.
102 +) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
103 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
104 the operation, and provides a more encapsulated way for external code
105 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
106 to use these functions rather than manually incrementing the counts.
109 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
110 This function was broken, as the check for a new client hello message
111 to handle SGC did not allow these large messages.
112 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
115 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
118 +) Add EVP test program.
121 +) Add symmetric cipher support to ENGINE. Expect the API to change!
124 +) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
125 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
126 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
127 These allow a CRL to be built without having to access X509_CRL fields
128 directly. Modify 'ca' application to use new functions.
131 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
132 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
135 *) Rework the configuration and shared library support for Tru64 Unix.
136 The configuration part makes use of modern compiler features and
137 still retains old compiler behavior for those that run older versions
138 of the OS. The shared library support part includes a variant that
139 uses the RPATH feature, and is available through the speciel
140 configuration target "alpha-cc-rpath", which will never be selected
142 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
144 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
145 with the same message size as in ssl3_get_certificate_request().
146 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
147 messages might inadvertently be reject as too long.
148 [Petr Lampa <lampa@fee.vutbr.cz>]
150 +) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
151 bug workarounds. Rollback attack detection is a security feature.
152 The problem will only arise on OpenSSL servers when TLSv1 is not
153 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
154 Software authors not wanting to support TLSv1 will have special reasons
155 for their choice and can explicitly enable this option.
156 [Bodo Moeller, Lutz Jaenicke]
158 +) Rationalise EVP so it can be extended: don't include a union of
159 cipher/digest structures, add init/cleanup functions. This also reduces
160 the number of header dependencies.
163 +) Make DES key schedule conform to the usual scheme, as well as
164 correcting its structure. This means that calls to DES functions
165 now have to pass a pointer to a des_key_schedule instead of a
166 plain des_key_schedule (which was actually always a pointer
170 +) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
173 *) Modified SSL library such that the verify_callback that has been set
174 specificly for an SSL object with SSL_set_verify() is actually being
175 used. Before the change, a verify_callback set with this function was
176 ignored and the verify_callback() set in the SSL_CTX at the time of
177 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
178 to allow the necessary settings.
181 +) Initial reduction of linker bloat: the use of some functions, such as
182 PEM causes large amounts of unused functions to be linked in due to
183 poor organisation. For example pem_all.c contains every PEM function
184 which has a knock on effect of linking in large amounts of (unused)
185 ASN1 code. Grouping together similar functions and splitting unrelated
186 functions prevents this.
189 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
190 explicitely to NULL, as at least on Solaris 8 this seems not always to be
191 done automatically (in contradiction to the requirements of the C
192 standard). This made problems when used from OpenSSH.
195 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
196 dh->length and always used
198 BN_rand_range(priv_key, dh->p).
200 BN_rand_range() is not necessary for Diffie-Hellman, and this
201 specific range makes Diffie-Hellman unnecessarily inefficient if
202 dh->length (recommended exponent length) is much smaller than the
203 length of dh->p. We could use BN_rand_range() if the order of
204 the subgroup was stored in the DH structure, but we only have
209 BN_rand(priv_key, l, ...)
211 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
217 RSA_eay_public_encrypt
218 RSA_eay_private_decrypt
219 RSA_eay_private_encrypt (signing)
220 RSA_eay_public_decrypt (signature verification)
222 (default implementations for RSA_public_encrypt,
223 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
224 always reject numbers >= n.
227 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
228 to synchronize access to 'locking_thread'. This is necessary on
229 systems where access to 'locking_thread' (an 'unsigned long'
230 variable) is not atomic.
233 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
234 *before* setting the 'crypto_lock_rand' flag. The previous code had
235 a race condition if 0 is a valid thread ID.
236 [Travis Vitek <vitek@roguewave.com>]
238 +) Cleanup of EVP macros.
241 +) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
245 +) Add initial OCSP responder support to ocsp application. The
246 revocation information is handled using the text based index
247 use by the ca application. The responder can either handle
248 requests generated internally, supplied in files (for example
249 via a CGI script) or using an internal minimal server.
252 +) Add configuration choices to get zlib compression for TLS.
255 +) Changes to Kerberos SSL for RFC 2712 compliance:
256 1. Implemented real KerberosWrapper, instead of just using
257 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
258 2. Implemented optional authenticator field of KerberosWrapper.
260 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
261 and authenticator structs; see crypto/krb5/.
263 Generalized Kerberos calls to support multiple Kerberos libraries.
264 [Vern Staats <staatsvr@asc.hpc.mil>,
265 Jeffrey Altman <jaltman@columbia.edu>
268 +) Cause 'openssl speed' to use fully hard-coded DSA keys as it
269 already does with RSA. testdsa.h now has 'priv_key/pub_key'
270 values for each of the key sizes rather than having just
271 parameters (and 'speed' generating keys each time).
274 -) OpenSSL 0.9.6b released [9 July 2001]
276 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
277 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
278 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
279 PRNG state recovery was possible based on the output of
280 one PRNG request appropriately sized to gain knowledge on
281 'md' followed by enough consecutive 1-byte PRNG requests
282 to traverse all of 'state'.
284 1. When updating 'md_local' (the current thread's copy of 'md')
285 during PRNG output generation, hash all of the previous
286 'md_local' value, not just the half used for PRNG output.
288 2. Make the number of bytes from 'state' included into the hash
289 independent from the number of PRNG bytes requested.
291 The first measure alone would be sufficient to avoid
292 Markku-Juhani's attack. (Actually it had never occurred
293 to me that the half of 'md_local' used for chaining was the
294 half from which PRNG output bytes were taken -- I had always
295 assumed that the secret half would be used.) The second
296 measure makes sure that additional data from 'state' is never
297 mixed into 'md_local' in small portions; this heuristically
298 further strengthens the PRNG.
301 +) Speed up EVP routines.
304 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
305 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
306 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
307 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
309 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
310 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
311 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
314 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
316 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
319 *) Fix crypto/bn/asm/mips3.s.
322 *) When only the key is given to "enc", the IV is undefined. Print out
323 an error message in this case.
326 +) Added the OS2-EMX target.
327 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
329 +) Rewrite apps to use NCONF routines instead of the old CONF. New functions
330 to support NCONF routines in extension code. New function CONF_set_nconf()
331 to allow functions which take an NCONF to also handle the old LHASH
332 structure: this means that the old CONF compatible routines can be
333 retained (in particular wrt extensions) without having to duplicate the
334 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
337 *) Handle special case when X509_NAME is empty in X509 printing routines.
340 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
341 positive and less than q.
344 +) Enhance the general user interface with mechanisms for inner control
345 and with pssibilities to have yes/no kind of prompts.
348 +) Change all calls to low level digest routines in the library and
349 applications to use EVP. Add missing calls to HMAC_cleanup() and
350 don't assume HMAC_CTX can be copied using memcpy().
351 [Verdon Walker <VWalker@novell.com>, Steve Henson]
353 +) Add the possibility to control engines through control names but with
354 arbitrary arguments instead of just a string.
355 Change the key loaders to take a UI_METHOD instead of a callback
356 function pointer. NOTE: this breaks binary compatibility with earlier
357 versions of OpenSSL [engine].
358 Addapt the nCipher code for these new conditions and add a card insertion
362 +) Enhance the general user interface with mechanisms to better support
363 dialog box interfaces, application-defined prompts, the possibility
364 to use defaults (for example default passwords from somewhere else)
365 and interrupts/cancelations.
368 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
369 used: it isn't thread safe and the add_lock_callback should handle
371 [Paul Rose <Paul.Rose@bridge.com>]
373 *) Verify that incoming data obeys the block size in
374 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
377 +) Tidy up PKCS#12 attribute handling. Add support for the CSP name
378 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
382 [Ulf Möller, Bodo Möller]
384 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
385 RSA encryption was accidentily removed in s3_srvr.c in OpenSSL 0.9.5
386 when fixing the server behaviour for backwards-compatible 'client
387 hello' messages. (Note that the attack is impractical against
388 SSL 3.0 and TLS 1.0 anyway because length and version checking
389 means that the probability of guessing a valid ciphertext is
390 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
393 Before 0.9.5, the countermeasure (hide the error by generating a
394 random 'decryption result') did not work properly because
395 ERR_clear_error() was missing, meaning that SSL_get_error() would
396 detect the supposedly ignored error.
398 Both problems are now fixed.
401 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
402 (previously it was 1024).
405 +) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
406 tidy up some unecessarily weird code in 'sk_new()').
407 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
409 +) Change the key loading routines for ENGINEs to use the same kind
410 callback (pem_password_cb) as all other routines that need this
414 *) Fix for compatibility mode trust settings: ignore trust settings
415 unless some valid trust or reject settings are present.
418 *) Fix for blowfish EVP: its a variable length cipher.
421 +) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
422 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
423 than this minimum value is recommended.
426 +) New random seeder for OpenVMS, using the system process statistics
427 that are easily reachable.
430 +) Windows apparently can't transparently handle global
431 variables defined in DLLs. Initialisations such as:
433 const ASN1_ITEM *it = &ASN1_INTEGER_it;
435 wont compile. This is used by the any applications that need to
436 delcare their own ASN1 modules. This was fixed by adding the option
437 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
438 needed for static libraries under Win32.
441 +) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
442 setting of purpose and trust fields. New X509_STORE trust and
443 purpose functions and tidy up setting in other SSL functions.
446 +) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
447 structure. These are inherited by X509_STORE_CTX when it is
448 initialised. This allows various defaults to be set in the
449 X509_STORE structure (such as flags for CRL checking and custom
450 purpose or trust settings) for functions which only use X509_STORE_CTX
451 internally such as S/MIME.
453 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
454 trust settings if they are not set in X509_STORE. This allows X509_STORE
455 purposes and trust (in S/MIME for example) to override any set by default.
457 Add command line options for CRL checking to smime, s_client and s_server
461 +) Initial CRL based revocation checking. If the CRL checking flag(s)
462 are set then the CRL is looked up in the X509_STORE structure and
463 its validity and signature checked, then if the certificate is found
464 in the CRL the verify fails with a revoked error.
466 Various new CRL related callbacks added to X509_STORE_CTX structure.
468 Command line options added to 'verify' application to support this.
470 This needs some additional work, such as being able to handle multiple
471 CRLs with different times, extension based lookup (rather than just
472 by subject name) and ultimately more complete V2 CRL extension
476 +) Add a general user interface API (crypto/ui/). This is designed
477 to replace things like des_read_password and friends (backward
478 compatibility functions using this new API are provided).
479 The purpose is to remove prompting functions from the DES code
480 section as well as provide for prompting through dialog boxes in
481 a window system and the like.
484 *) Fix various bugs related to DSA S/MIME verification. Handle missing
485 parameters in DSA public key structures and return an error in the
486 DSA routines if parameters are absent.
489 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
490 in the current directory if neither $RANDFILE nor $HOME was set.
491 RAND_file_name() in 0.9.6a returned NULL in this case. This has
492 caused some confusion to Windows users who haven't defined $HOME.
493 Thus RAND_file_name() is changed again: e_os.h can define a
494 DEFAULT_HOME, which will be used if $HOME is not set.
495 For Windows, we use "C:"; on other platforms, we still require
496 environment variables.
498 +) Add "ex_data" support to ENGINE so implementations can add state at a
499 per-structure level rather than having to store it globally.
502 +) Make it possible for ENGINE structures to be copied when retrieved by
503 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
504 This causes the "original" ENGINE structure to act like a template,
505 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
506 operational state can be localised to each ENGINE structure, despite the
507 fact they all share the same "methods". New ENGINE structures returned in
508 this case have no functional references and the return value is the single
509 structural reference. This matches the single structural reference returned
510 by ENGINE_by_id() normally, when it is incremented on the pre-existing
514 +) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
515 needs to match any other type at all we need to manually clear the
519 +) Changes to the "openssl engine" utility to include;
520 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
521 about an ENGINE's available control commands.
522 - executing control commands from command line arguments using the
523 '-pre' and '-post' switches. '-post' is only used if '-t' is
524 specified and the ENGINE is successfully initialised. The syntax for
525 the individual commands are colon-separated, for example;
526 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
529 +) New dynamic control command support for ENGINEs. ENGINEs can now
530 declare their own commands (numbers), names (strings), descriptions,
531 and input types for run-time discovery by calling applications. A
532 subset of these commands are implicitly classed as "executable"
533 depending on their input type, and only these can be invoked through
534 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
535 can be based on user input, config files, etc). The distinction is
536 that "executable" commands cannot return anything other than a boolean
537 result and can only support numeric or string input, whereas some
538 discoverable commands may only be for direct use through
539 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
540 pointers, or other custom uses. The "executable" commands are to
541 support parameterisations of ENGINE behaviour that can be
542 unambiguously defined by ENGINEs and used consistently across any
543 OpenSSL-based application. Commands have been added to all the
544 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
545 control over shared-library paths without source code alterations.
548 +) Changed all ENGINE implementations to dynamically allocate their
549 ENGINEs rather than declaring them statically. Apart from this being
550 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
551 this also allows the implementations to compile without using the
552 internal engine_int.h header.
555 +) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
556 'const' value. Any code that should be able to modify a RAND_METHOD
557 should already have non-const pointers to it (ie. they should only
558 modify their own ones).
561 +) Made a variety of little tweaks to the ENGINE code.
562 - "atalla" and "ubsec" string definitions were moved from header files
563 to C code. "nuron" string definitions were placed in variables
564 rather than hard-coded - allowing parameterisation of these values
565 later on via ctrl() commands.
566 - Removed unused "#if 0"'d code.
567 - Fixed engine list iteration code so it uses ENGINE_free() to release
568 structural references.
569 - Constified the RAND_METHOD element of ENGINE structures.
570 - Constified various get/set functions as appropriate and added
571 missing functions (including a catch-all ENGINE_cpy that duplicates
572 all ENGINE values onto a new ENGINE except reference counts/state).
573 - Removed NULL parameter checks in get/set functions. Setting a method
574 or function to NULL is a way of cancelling out a previously set
575 value. Passing a NULL ENGINE parameter is just plain stupid anyway
576 and doesn't justify the extra error symbols and code.
577 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
578 flags from engine_int.h to engine.h.
579 - Changed prototypes for ENGINE handler functions (init(), finish(),
580 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
583 *) Move 'if (!initialized) RAND_poll()' into regions protected by
584 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
585 having multiple threads call RAND_poll() concurrently.
588 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
589 combination of a flag and a thread ID variable.
590 Otherwise while one thread is in ssleay_rand_bytes (which sets the
591 flag), *other* threads can enter ssleay_add_bytes without obeying
592 the CRYPTO_LOCK_RAND lock (and may even illegaly release the lock
593 that they do not hold after the first thread unsets add_do_not_lock).
596 +) Implement binary inversion algorithm for BN_mod_inverse in addition
597 to the algorithm using long divison. The binary algorithm can be
598 used only if the modulus is odd. On 32-bit systems, it is faster
599 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
600 roughly 5-15% for 256-bit moduli), so we use it only for moduli
601 up to 450 bits. In 64-bit environments, the binary algorithm
602 appears to be advantageous for much longer moduli; here we use it
603 for moduli up to 2048 bits.
606 *) Change bctest again: '-x' expressions are not available in all
610 -) OpenSSL 0.9.6a released [5 Apr 2001]
612 *) Fix a couple of memory leaks in PKCS7_dataDecode()
613 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
615 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
616 the default extension for executables, if any. Also, make the perl
617 scripts that use symlink() to test if it really exists and use "cp"
618 if it doesn't. All this made OpenSSL compilable and installable in
622 +) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
623 could not support the combine flag in choice fields.
626 -) Fix for asn1_GetSequence() for indefinite length constructed data.
627 If SEQUENCE is length is indefinite just set c->slen to the total
628 amount of data available.
629 [Steve Henson, reported by shige@FreeBSD.org]
630 [This change does not apply to 0.9.7.]
632 *) Change bctest to avoid here-documents inside command substitution
633 (workaround for FreeBSD /bin/sh bug).
634 For compatibility with Ultrix, avoid shell functions (introduced
635 in the bctest version that searches along $PATH).
638 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
639 with des_encrypt() defined on some operating systems, like Solaris
643 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
644 On the Importance of Eliminating Errors in Cryptographic
645 Computations, J. Cryptology 14 (2001) 2, 101-119,
646 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
649 *) MIPS assembler BIGNUM division bug fix.
652 *) Disabled incorrect Alpha assembler code.
655 -) Fix PKCS#7 decode routines so they correctly update the length
656 after reading an EOC for the EXPLICIT tag.
658 [This change does not apply to 0.9.7.]
660 *) Fix bug in PKCS#12 key generation routines. This was triggered
661 if a 3DES key was generated with a 0 initial byte. Include
662 PKCS12_BROKEN_KEYGEN compilation option to retain the old
663 (but broken) behaviour.
666 *) Enhance bctest to search for a working bc along $PATH and print
668 [Tim Rice <tim@multitalents.net> via Richard Levitte]
670 +) Add a 'copy_extensions' option to the 'ca' utility. This copies
671 extensions from a certificate request to the certificate.
674 +) Allow multiple 'certopt' and 'nameopt' options to be separated
675 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
676 file: this allows the display of the certificate about to be
677 signed to be customised, to allow certain fields to be included
678 or excluded and extension details. The old system didn't display
679 multicharacter strings properly, omitted fields not in the policy
680 and couldn't display additional details such as extensions.
683 *) Fix memory leaks in err.c: free err_data string if necessary;
684 don't write to the wrong index in ERR_set_error_data.
687 +) Function EC_POINTs_mul for simultaneous scalar multiplication
688 of an arbitrary number of elliptic curve points, optionally
689 including the generator defined for the EC_GROUP.
690 EC_POINT_mul is a simple wrapper function for the typical case
691 that the point list has just one item (besides the optional
695 +) First EC_METHODs for curves over GF(p):
697 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
698 operations and provides various method functions that can also
699 operate with faster implementations of modular arithmetic.
701 EC_GFp_mont_method() reuses most functions that are part of
702 EC_GFp_simple_method, but uses Montgomery arithmetic.
704 [Bodo Moeller; point addition and point doubling
705 implementation directly derived from source code provided by
706 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
708 +) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
711 Curves are EC_GROUP objects (with an optional group generator)
712 based on EC_METHODs that are built into the library.
714 Points are EC_POINT objects based on EC_GROUP objects.
716 Most of the framework would be able to handle curves over arbitrary
717 finite fields, but as there are no obvious types for fields other
718 than GF(p), some functions are limited to that for now.
721 +) Add the -HTTP option to s_server. It is similar to -WWW, but requires
722 that the file contains a complete HTTP response.
725 +) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
726 change the def and num file printf format specifier from "%-40sXXX"
727 to "%-39s XXX". The latter will always guarantee a space after the
728 field while the former will cause them to run together if the field
729 is 40 of more characters long.
732 +) Constify the cipher and digest 'method' functions and structures
733 and modify related functions to take constant EVP_MD and EVP_CIPHER
737 *) Implement ssl23_peek (analogous to ssl23_read), which previously
741 *) Replace rdtsc with _emit statements for VC++ version 5.
742 [Jeremy Cooper <jeremy@baymoo.org>]
744 +) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
745 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
748 +) Modify EVP_Digest*() routines so they now return values. Although the
749 internal software routines can never fail additional hardware versions
753 +) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
755 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
756 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
759 ERR_R_NESTED_ASN1_ERROR
761 ERR_R_MISSING_ASN1_EOS
762 were 4 .. 9, conflicting with
763 ERR_LIB_RSA (= ERR_R_RSA_LIB)
765 ERR_LIB_PEM (= ERR_R_PEM_LIB).
766 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
768 Add new error code 'ERR_R_INTERNAL_ERROR'.
771 +) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
775 +) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
776 sets the subject name for a new request or supersedes the
777 subject name in a given request. Formats that can be parsed are
778 'CN=Some Name, OU=myOU, C=IT'
780 'CN=Some Name/OU=myOU/C=IT'.
782 Add options '-batch' and '-verbose' to 'openssl req'.
783 [Massimiliano Pala <madwolf@hackmasters.net>]
785 *) Make it possible to reuse SSLv2 sessions.
788 +) Introduce the possibility to access global variables through
789 functions on platform were that's the best way to handle exporting
790 global variables in shared libraries. To enable this functionality,
791 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
792 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
793 is normally done by Configure or something similar).
795 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
796 in the source file (foo.c) like this:
798 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
799 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
801 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
802 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
804 OPENSSL_DECLARE_GLOBAL(int,foo);
805 #define foo OPENSSL_GLOBAL_REF(foo)
806 OPENSSL_DECLARE_GLOBAL(double,bar);
807 #define bar OPENSSL_GLOBAL_REF(bar)
809 The #defines are very important, and therefore so is including the
810 header file everywere where the defined globals are used.
812 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
813 of ASN.1 items, but that structure is a bt different.
815 The largest change is in util/mkdef.pl which has been enhanced with
816 better and easier to understand logic to choose which symbols should
817 go into the Windows .def files as well as a number of fixes and code
818 cleanup (among others, algorithm keywords are now sorted
819 lexicographically to avoid constant rewrites).
822 *) In copy_email() check for >= 0 as a return value for
823 X509_NAME_get_index_by_NID() since 0 is a valid index.
824 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
826 +) In BN_div() keep a copy of the sign of 'num' before writing the
827 result to 'rm' because if rm==num the value will be overwritten
828 and produce the wrong result if 'num' is negative: this caused
829 problems with BN_mod() and BN_nnmod().
832 +) Function OCSP_request_verify(). This checks the signature on an
833 OCSP request and verifies the signer certificate. The signer
834 certificate is just checked for a generic purpose and OCSP request
838 +) Add OCSP_check_validity() function to check the validity of OCSP
839 responses. OCSP responses are prepared in real time and may only
840 be a few seconds old. Simply checking that the current time lies
841 between thisUpdate and nextUpdate max reject otherwise valid responses
842 caused by either OCSP responder or client clock innacuracy. Instead
843 we allow thisUpdate and nextUpdate to fall within a certain period of
844 the current time. The age of the response can also optionally be
845 checked. Two new options -validity_period and -status_age added to
849 +) If signature or public key algorithm is unrecognized print out its
850 OID rather that just UNKOWN.
853 *) Avoid coredump with unsupported or invalid public keys by checking if
854 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
855 PKCS7_verify() fails with non detached data.
858 +) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
859 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
860 ID to be generated from the issuer certificate alone which can then be
861 passed to OCSP_id_issuer_cmp().
864 +) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
865 ASN1 modules to export functions returning ASN1_ITEM pointers
866 instead of the ASN1_ITEM structures themselves. This adds several
867 new macros which allow the underlying ASN1 function/structure to
868 be accessed transparently. As a result code should not use ASN1_ITEM
869 references directly (such as &X509_it) but instead use the relevant
870 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
871 use of the new ASN1 code on platforms where exporting structures
872 is problematical (for example in shared libraries) but exporting
873 functions returning pointers to structures is not.
876 +) Add support for overriding the generation of SSL/TLS session IDs.
877 These callbacks can be registered either in an SSL_CTX or per SSL.
878 The purpose of this is to allow applications to control, if they wish,
879 the arbitrary values chosen for use as session IDs, particularly as it
880 can be useful for session caching in multiple-server environments. A
881 command-line switch for testing this (and any client code that wishes
882 to use such a feature) has been added to "s_server".
883 [Geoff Thorpe, Lutz Jaenicke]
885 +) Modify mkdef.pl to recognise and parse prprocessor conditionals
886 of the form '#if defined(...) || defined(...) || ...' and
887 '#if !defined(...) && !defined(...) && ...'. This also avoids
888 the growing number of special cases it was previously handling.
891 *) Don't use getenv in library functions when run as setuid/setgid.
892 New function OPENSSL_issetugid().
895 +) Make all configuration macros available for application by making
896 sure they are available in opensslconf.h, by giving them names starting
897 with "OPENSSL_" to avoid conflicts with other packages and by making
898 sure e_os2.h will cover all platform-specific cases together with
900 Additionally, it is now possible to define configuration/platform-
901 specific names (called "system identities"). In the C code, these
902 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
903 macro with the name beginning with "OPENSSL_SYS_", which is determined
904 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
908 +) New option -set_serial to 'req' and 'x509' this allows the serial
909 number to use to be specified on the command line. Previously self
910 signed certificates were hard coded with serial number 0 and the
911 CA options of 'x509' had to use a serial number in a file which was
915 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
916 due to incorrect handling of multi-threading:
918 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
920 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
922 3. Count how many times MemCheck_off() has been called so that
923 nested use can be treated correctly. This also avoids
924 inband-signalling in the previous code (which relied on the
925 assumption that thread ID 0 is impossible).
928 +) New options to 'ca' utility to support V2 CRL entry extensions.
929 Currently CRL reason, invalidity date and hold instruction are
930 supported. Add new CRL extensions to V3 code and some new objects.
933 *) Add "-rand" option also to s_client and s_server.
936 +) New function EVP_CIPHER_CTX_set_padding() this is used to
937 disable standard block padding (aka PKCS#5 padding) in the EVP
938 API, which was previously mandatory. This means that the data is
939 not padded in any way and so the total length much be a multiple
940 of the block size, otherwise an error occurs.
943 +) Initial (incomplete) OCSP SSL support.
946 *) Fix CPU detection on Irix 6.x.
947 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
948 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
950 +) New function OCSP_parse_url(). This splits up a URL into its host,
951 port and path components: primarily to parse OCSP URLs. New -url
952 option to ocsp utility.
955 +) New nonce behavior. The return value of OCSP_check_nonce() now
956 reflects the various checks performed. Applications can decide
957 whether to tolerate certain situations such as an absent nonce
958 in a response when one was present in a request: the ocsp application
959 just prints out a warning. New function OCSP_add1_basic_nonce()
960 this is to allow responders to include a nonce in a response even if
961 the request is nonce-less.
964 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
967 [This change does not apply to 0.9.7.]
969 *) Use the cached encoding of an X509_NAME structure rather than
970 copying it. This is apparently the reason for the libsafe "errors"
971 but the code is actually correct.
974 +) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
975 skipped when using openssl x509 multiple times on a single input file,
976 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
979 +) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
980 set string type: to handle setting ASN1_TIME structures. Fix ca
981 utility to correctly initialize revocation date of CRLs.
984 +) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
985 the clients preferred ciphersuites and rather use its own preferences.
986 Should help to work around M$ SGC (Server Gated Cryptography) bug in
987 Internet Explorer by ensuring unchanged hash method during stepup.
988 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
991 +) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
992 to aes and add a new 'exist' option to print out symbols that don't
996 +) Additional options to ocsp utility to allow flags to be set and
997 additional certificates supplied.
1000 +) Add the option -VAfile to 'openssl ocsp', so the user can give the
1001 OCSP client a number of certificate to only verify the response
1005 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
1006 Bleichenbacher's DSA attack.
1007 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
1008 to be set and top=0 forces the highest bit to be set; top=-1 is new
1009 and leaves the highest bit random.
1010 [Ulf Moeller, Bodo Moeller]
1012 +) Update Rijndael code to version 3.0 and change EVP AES ciphers to
1013 handle the new API. Currently only ECB, CBC modes supported. Add new
1014 AES OIDs. Add TLS AES ciphersuites as described in the "AES Ciphersuites
1015 for TLS" draft-ietf-tls-ciphersuite-03.txt.
1016 [Ben Laurie, Steve Henson]
1018 *) In the NCONF_...-based implementations for CONF_... queries
1019 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
1020 a temporary CONF structure with the data component set to NULL
1021 (which gives segmentation faults in lh_retrieve).
1022 Instead, use NULL for the CONF pointer in CONF_get_string and
1023 CONF_get_number (which may use environment variables) and directly
1024 return NULL from CONF_get_section.
1027 *) Fix potential buffer overrun for EBCDIC.
1030 +) New function OCSP_copy_nonce() to copy nonce value (if present) from
1031 request to response.
1034 +) Functions for OCSP responders. OCSP_request_onereq_count(),
1035 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
1036 extract information from a certificate request. OCSP_response_create()
1037 creates a response and optionally adds a basic response structure.
1038 OCSP_basic_add1_status() adds a complete single response to a basic
1039 reponse and returns the OCSP_SINGLERESP structure just added (to allow
1040 extensions to be included for example). OCSP_basic_add1_cert() adds a
1041 certificate to a basic response and OCSP_basic_sign() signs a basic
1042 response with various flags. New helper functions ASN1_TIME_check()
1043 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
1044 (converts ASN1_TIME to GeneralizedTime).
1047 +) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
1048 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
1049 structure from a certificate. X509_pubkey_digest() digests tha public_key
1050 contents: this is used in various key identifiers.
1053 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
1054 keyUsage if basicConstraints absent for a CA.
1057 *) Make SMIME_write_PKCS7() write mail header values with a format that
1058 is more generally accepted (no spaces before the semicolon), since
1059 some programs can't parse those values properly otherwise. Also make
1060 sure BIO's that break lines after each write do not create invalid
1064 +) Make sk_sort() tolerate a NULL argument.
1065 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
1067 +) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
1068 passed by the function are trusted implicitly. If any of them signed the
1069 reponse then it is assumed to be valid and is not verified.
1072 -) Make the CRL encoding routines work with empty SEQUENCE OF. The
1073 macros previously used would not encode an empty SEQUENCE OF
1074 and break the signature.
1076 [This change does not apply to 0.9.7.]
1078 *) Zero the premaster secret after deriving the master secret in
1082 +) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
1083 to data. This was previously part of the PKCS7 ASN1 code. This
1084 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
1085 [Steve Henson, reported by Kenneth R. Robinette
1086 <support@securenetterm.com>]
1088 +) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
1089 routines: without these tracing memory leaks is very painful.
1090 Fix leaks in PKCS12 and PKCS7 routines.
1093 *) Add some EVP_add_digest_alias registrations (as found in
1094 OpenSSL_add_all_digests()) to SSL_library_init()
1095 aka OpenSSL_add_ssl_algorithms(). This provides improved
1096 compatibility with peers using X.509 certificates
1097 with unconventional AlgorithmIdentifier OIDs.
1100 *) Fix for Irix with NO_ASM.
1101 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
1103 *) ./config script fixes.
1104 [Ulf Moeller, Richard Levitte]
1106 +) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
1107 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
1108 effectively meant GeneralizedTime would never be used. Now it
1109 is initialised to -1 but X509_time_adj() now has to check the value
1110 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
1111 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
1112 [Steve Henson, reported by Kenneth R. Robinette
1113 <support@securenetterm.com>]
1115 +) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
1116 result in a zero length in the ASN1_INTEGER structure which was
1117 not consistent with the structure when d2i_ASN1_INTEGER() was used
1118 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
1119 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
1120 where it did not print out a minus for negative ASN1_INTEGER.
1123 *) Fix 'openssl passwd -1'.
1126 +) Add summary printout to ocsp utility. The various functions which
1127 convert status values to strings have been renamed to:
1128 OCSP_response_status_str(), OCSP_cert_status_str() and
1129 OCSP_crl_reason_str() and are no longer static. New options
1130 to verify nonce values and to disable verification. OCSP response
1131 printout format cleaned up.
1134 +) Add additional OCSP certificate checks. These are those specified
1135 in RFC2560. This consists of two separate checks: the CA of the
1136 certificate being checked must either be the OCSP signer certificate
1137 or the issuer of the OCSP signer certificate. In the latter case the
1138 OCSP signer certificate must contain the OCSP signing extended key
1139 usage. This check is performed by attempting to match the OCSP
1140 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
1141 in the OCSP_CERTID structures of the response.
1144 +) Initial OCSP certificate verification added to OCSP_basic_verify()
1145 and related routines. This uses the standard OpenSSL certificate
1146 verify routines to perform initial checks (just CA validity) and
1147 to obtain the certificate chain. Then additional checks will be
1148 performed on the chain. Currently the root CA is checked to see
1149 if it is explicitly trusted for OCSP signing. This is used to set
1150 a root CA as a global signing root: that is any certificate that
1151 chains to that CA is an acceptable OCSP signing certificate.
1154 +) New '-extfile ...' option to 'openssl ca' for reading X.509v3
1155 extensions from a separate configuration file.
1156 As when reading extensions from the main configuration file,
1157 the '-extensions ...' option may be used for specifying the
1159 [Massimiliano Pala <madwolf@comune.modena.it>]
1161 *) Change PKCS12_key_gen_asc() so it can cope with non null
1162 terminated strings whose length is passed in the passlen
1163 parameter, for example from PEM callbacks. This was done
1164 by adding an extra length parameter to asc2uni().
1165 [Steve Henson, reported by <oddissey@samsung.co.kr>]
1167 +) New OCSP utility. Allows OCSP requests to be generated or
1168 read. The request can be sent to a responder and the output
1169 parsed, outputed or printed in text form. Not complete yet:
1170 still needs to check the OCSP response validity.
1173 +) New subcommands for 'openssl ca':
1174 'openssl ca -status <serial>' prints the status of the cert with
1175 the given serial number (according to the index file).
1176 'openssl ca -updatedb' updates the expiry status of certificates
1178 [Massimiliano Pala <madwolf@comune.modena.it>]
1180 +) New '-newreq-nodes' command option to CA.pl. This is like
1181 '-newreq', but calls 'openssl req' with the '-nodes' option
1182 so that the resulting key is not encrypted.
1183 [Damien Miller <djm@mindrot.org>]
1185 +) New configuration for the GNU Hurd.
1186 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
1188 +) Initial code to implement OCSP basic response verify. This
1189 is currently incomplete. Currently just finds the signer's
1190 certificate and verifies the signature on the response.
1193 +) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
1194 value of OPENSSLDIR. This is available via the new '-d' option
1195 to 'openssl version', and is also included in 'openssl version -a'.
1198 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
1199 call failed, free the DSA structure.
1202 +) Allowing defining memory allocation callbacks that will be given
1203 file name and line number information in additional arguments
1204 (a const char* and an int). The basic functionality remains, as
1205 well as the original possibility to just replace malloc(),
1206 realloc() and free() by functions that do not know about these
1207 additional arguments. To register and find out the current
1208 settings for extended allocation functions, the following
1209 functions are provided:
1211 CRYPTO_set_mem_ex_functions
1212 CRYPTO_set_locked_mem_ex_functions
1213 CRYPTO_get_mem_ex_functions
1214 CRYPTO_get_locked_mem_ex_functions
1216 These work the same way as CRYPTO_set_mem_functions and friends.
1217 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
1218 extended allocation function is enabled.
1219 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
1220 a conventional allocation function is enabled.
1221 [Richard Levitte, Bodo Moeller]
1223 *) Fix to uni2asc() to cope with zero length Unicode strings.
1224 These are present in some PKCS#12 files.
1227 +) Finish off removing the remaining LHASH function pointer casts.
1228 There should no longer be any prototype-casting required when using
1229 the LHASH abstraction, and any casts that remain are "bugs". See
1230 the callback types and macros at the head of lhash.h for details
1231 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
1234 +) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
1235 If /dev/[u]random devices are not available or do not return enough
1236 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
1238 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
1239 /etc/entropy will be queried once each in this sequence, quering stops
1240 when enough entropy was collected without querying more sockets.
1243 +) Change the Unix RAND_poll() variant to be able to poll several
1244 random devices, as specified by DEVRANDOM, until a sufficient amount
1245 of data has been collected. We spend at most 10 ms on each file
1246 (select timeout) and read in non-blocking mode. DEVRANDOM now
1247 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
1248 (previously it was just the string "/dev/urandom"), so on typical
1249 platforms the 10 ms delay will never occur.
1250 Also separate out the Unix variant to its own file, rand_unix.c.
1251 For VMS, there's a currently-empty rand_vms.c.
1254 +) Move OCSP client related routines to ocsp_cl.c. These
1255 provide utility functions which an application needing
1256 to issue a request to an OCSP responder and analyse the
1257 response will typically need: as opposed to those which an
1258 OCSP responder itself would need which will be added later.
1260 OCSP_request_sign() signs an OCSP request with an API similar
1261 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
1262 response. OCSP_response_get1_basic() extracts basic response
1263 from response. OCSP_resp_find_status(): finds and extracts status
1264 information from an OCSP_CERTID structure (which will be created
1265 when the request structure is built). These are built from lower
1266 level functions which work on OCSP_SINGLERESP structures but
1267 wont normally be used unless the application wishes to examine
1268 extensions in the OCSP response for example.
1270 Replace nonce routines with a pair of functions.
1271 OCSP_request_add1_nonce() adds a nonce value and optionally
1272 generates a random value. OCSP_check_nonce() checks the
1273 validity of the nonce in an OCSP response.
1276 +) Change function OCSP_request_add() to OCSP_request_add0_id().
1277 This doesn't copy the supplied OCSP_CERTID and avoids the
1278 need to free up the newly created id. Change return type
1279 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
1280 This can then be used to add extensions to the request.
1281 Deleted OCSP_request_new(), since most of its functionality
1282 is now in OCSP_REQUEST_new() (and the case insensitive name
1283 clash) apart from the ability to set the request name which
1284 will be added elsewhere.
1287 +) Update OCSP API. Remove obsolete extensions argument from
1288 various functions. Extensions are now handled using the new
1289 OCSP extension code. New simple OCSP HTTP function which
1290 can be used to send requests and parse the response.
1293 +) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
1294 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
1295 uses the special reorder version of SET OF to sort the attributes
1296 and reorder them to match the encoded order. This resolves a long
1297 standing problem: a verify on a PKCS7 structure just after signing
1298 it used to fail because the attribute order did not match the
1299 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
1300 it uses the received order. This is necessary to tolerate some broken
1301 software that does not order SET OF. This is handled by encoding
1302 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
1303 to produce the required SET OF.
1306 +) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
1307 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
1308 files to get correct declarations of the ASN.1 item variables.
1311 +) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
1312 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
1313 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
1314 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
1315 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
1316 ASN1_ITEM and no wrapper functions.
1319 +) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
1320 replace the old function pointer based I/O routines. Change most of
1321 the *_d2i_bio() and *_d2i_fp() functions to use these.
1324 +) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
1325 lines, recognice more "algorithms" that can be deselected, and make
1326 it complain about algorithm deselection that isn't recognised.
1329 +) New ASN1 functions to handle dup, sign, verify, digest, pack and
1330 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
1331 to use new functions. Add NO_ASN1_OLD which can be set to remove
1332 some old style ASN1 functions: this can be used to determine if old
1333 code will still work when these eventually go away.
1336 +) New extension functions for OCSP structures, these follow the
1337 same conventions as certificates and CRLs.
1340 +) New function X509V3_add1_i2d(). This automatically encodes and
1341 adds an extension. Its behaviour can be customised with various
1342 flags to append, replace or delete. Various wrappers added for
1343 certifcates and CRLs.
1346 +) Fix to avoid calling the underlying ASN1 print routine when
1347 an extension cannot be parsed. Correct a typo in the
1348 OCSP_SERVICELOC extension. Tidy up print OCSP format.
1351 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
1352 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
1353 when writing a 32767 byte record.
1354 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
1356 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
1357 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
1359 (RSA objects have a reference count access to which is protected
1360 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
1361 so they are meant to be shared between threads.)
1362 [Bodo Moeller, Geoff Thorpe; original patch submitted by
1363 "Reddie, Steven" <Steven.Reddie@ca.com>]
1365 +) Make mkdef.pl parse some of the ASN1 macros and add apropriate
1366 entries for variables.
1369 *) Fix a deadlock in CRYPTO_mem_leaks().
1372 +) Add functionality to apps/openssl.c for detecting locking
1373 problems: As the program is single-threaded, all we have
1374 to do is register a locking callback using an array for
1375 storing which locks are currently held by the program.
1378 +) Use a lock around the call to CRYPTO_get_ex_new_index() in
1379 SSL_get_ex_data_X509_STORE_idx(), which is used in
1380 ssl_verify_cert_chain() and thus can be called at any time
1381 during TLS/SSL handshakes so that thread-safety is essential.
1382 Unfortunately, the ex_data design is not at all suited
1383 for multi-threaded use, so it probably should be abolished.
1386 +) Added Broadcom "ubsec" ENGINE to OpenSSL.
1387 [Broadcom, tweaked and integrated by Geoff Thorpe]
1389 +) Move common extension printing code to new function
1390 X509V3_print_extensions(). Reorganise OCSP print routines and
1391 implement some needed OCSP ASN1 functions. Add OCSP extensions.
1394 +) New function X509_signature_print() to remove duplication in some
1398 +) Add a special meaning when SET OF and SEQUENCE OF flags are both
1399 set (this was treated exactly the same as SET OF previously). This
1400 is used to reorder the STACK representing the structure to match the
1401 encoding. This will be used to get round a problem where a PKCS7
1402 structure which was signed could not be verified because the STACK
1403 order did not reflect the encoded order.
1406 +) Reimplement the OCSP ASN1 module using the new code.
1409 +) Update the X509V3 code to permit the use of an ASN1_ITEM structure
1410 for its ASN1 operations. The old style function pointers still exist
1411 for now but they will eventually go away.
1414 +) Merge in replacement ASN1 code from the ASN1 branch. This almost
1415 completely replaces the old ASN1 functionality with a table driven
1416 encoder and decoder which interprets an ASN1_ITEM structure describing
1417 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
1418 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
1419 has also been converted to the new form.
1422 +) Change BN_mod_exp_recp so that negative moduli are tolerated
1423 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
1424 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
1425 for negative moduli.
1428 +) Fix BN_uadd and BN_usub: Always return non-negative results instead
1429 of not touching the result's sign bit.
1432 +) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
1436 +) Changed the LHASH code to use prototypes for callbacks, and created
1437 macros to declare and implement thin (optionally static) functions
1438 that provide type-safety and avoid function pointer casting for the
1439 type-specific callbacks.
1442 *) Use better test patterns in bntest.
1445 +) Added Kerberos Cipher Suites to be used with TLS, as written in
1447 [Veers Staats <staatsvr@asc.hpc.mil>,
1448 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
1450 *) rand_win.c fix for Borland C.
1453 *) BN_rshift bugfix for n == 0.
1456 +) Reformat the FAQ so the different questions and answers can be divided
1457 in sections depending on the subject.
1460 +) Have the zlib compression code load ZLIB.DLL dynamically under
1464 +) New function BN_mod_sqrt for computing square roots modulo a prime
1465 (using the probabilistic Tonelli-Shanks algorithm unless
1466 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
1467 be handled deterministically).
1468 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
1470 *) Add a 'bctest' script that checks for some known 'bc' bugs
1471 so that 'make test' does not abort just because 'bc' is broken.
1474 *) Store verify_result within SSL_SESSION also for client side to
1475 avoid potential security hole. (Re-used sessions on the client side
1476 always resulted in verify_result==X509_V_OK, not using the original
1477 result of the server certificate verification.)
1480 +) Make BN_mod_inverse faster by explicitly handling small quotients
1481 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
1482 512 bits], about 30% for larger ones [1024 or 2048 bits].)
1485 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
1486 SSL3_RT_APPLICATION_DATA, return 0.
1487 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
1491 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
1492 releases, have been re-implemented by renaming the previous
1493 implementations of ssl2_read and ssl3_read to ssl2_read_internal
1494 and ssl3_read_internal, respectively, and adding 'peek' parameters
1495 to them. The new ssl[23]_{read,peek} functions are calls to
1496 ssl[23]_read_internal with the 'peek' flag set appropriately.
1497 A 'peek' parameter has also been added to ssl3_read_bytes, which
1498 does the actual work for ssl3_read_internal.
1501 +) New function BN_kronecker.
1504 +) Fix BN_gcd so that it works on negative inputs; the result is
1505 positive unless both parameters are zero.
1506 Previously something reasonably close to an infinite loop was
1507 possible because numbers could be growing instead of shrinking
1508 in the implementation of Euclid's algorithm.
1511 +) Fix BN_is_word() and BN_is_one() macros to take into account the
1512 sign of the number in question.
1514 Fix BN_is_word(a,w) to work correctly for w == 0.
1516 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
1517 because its test if the absolute value of 'a' equals 'w'.
1518 Note that BN_abs_is_word does *not* handle w == 0 reliably;
1519 it exists mostly for use in the implementations of BN_is_zero(),
1520 BN_is_one(), and BN_is_word().
1523 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
1524 the method-specific "init()" handler. Also clean up ex_data after
1525 calling the method-specific "finish()" handler. Previously, this was
1526 happening the other way round.
1529 +) New function BN_swap.
1532 +) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
1533 the exponentiation functions are more likely to produce reasonable
1534 results on negative inputs.
1537 +) Change BN_mod_mul so that the result is always non-negative.
1538 Previously, it could be negative if one of the factors was negative;
1539 I don't think anyone really wanted that behaviour.
1542 +) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
1543 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
1544 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
1545 and add new functions:
1554 BN_mod_lshift1_quick
1558 These functions always generate non-negative results.
1560 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
1561 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
1563 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
1564 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
1565 be reduced modulo m.
1566 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
1568 +) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
1569 was actually never needed) and in BN_mul(). The removal in BN_mul()
1570 required a small change in bn_mul_part_recursive() and the addition
1571 of the functions bn_cmp_part_words(), bn_sub_part_words() and
1572 bn_add_part_words(), which do the same thing as bn_cmp_words(),
1573 bn_sub_words() and bn_add_words() except they take arrays with
1577 +) In 'openssl passwd', verify passwords read from the terminal
1578 unless the '-salt' option is used (which usually means that
1579 verification would just waste user's time since the resulting
1580 hash is going to be compared with some given password hash)
1581 or the new '-noverify' option is used.
1583 This is an incompatible change, but it does not affect
1584 non-interactive use of 'openssl passwd' (passwords on the command
1585 line, '-stdin' option, '-in ...' option) and thus should not
1589 +) Remove all references to RSAref, since there's no more need for it.
1592 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
1593 The previous value, 12, was not always sufficient for BN_mod_exp().
1596 +) Make DSO load along a path given through an environment variable
1597 (SHLIB_PATH) with shl_load().
1600 +) Constify the ENGINE code as a result of BIGNUM constification.
1601 Also constify the RSA code and most things related to it. In a
1602 few places, most notable in the depth of the ASN.1 code, ugly
1603 casts back to non-const were required (to be solved at a later
1607 +) Make it so the openssl application has all engines loaded by default.
1610 +) Constify the BIGNUM routines a little more.
1613 *) Make sure that shared libraries get the internal name engine with
1614 the full version number and not just 0. This should mark the
1615 shared libraries as not backward compatible. Of course, this should
1616 be changed again when we can guarantee backward binary compatibility.
1619 +) Add the following functions:
1621 ENGINE_load_cswift()
1623 ENGINE_load_atalla()
1625 ENGINE_load_builtin_engines()
1627 That way, an application can itself choose if external engines that
1628 are built-in in OpenSSL shall ever be used or not. The benefit is
1629 that applications won't have to be linked with libdl or other dso
1630 libraries unless it's really needed.
1632 Changed 'openssl engine' to load all engines on demand.
1633 Changed the engine header files to avoid the duplication of some
1634 declarations (they differed!).
1637 +) 'openssl engine' can now list capabilities.
1640 +) Better error reporting in 'openssl engine'.
1643 +) Never call load_dh_param(NULL) in s_server.
1646 +) Add engine application. It can currently list engines by name and
1647 identity, and test if they are actually available.
1650 +) Add support for shared libraries under Irix.
1651 [Albert Chin-A-Young <china@thewrittenword.com>]
1653 +) Improve RPM specification file by forcing symbolic linking and making
1654 sure the installed documentation is also owned by root.root.
1655 [Damien Miller <djm@mindrot.org>]
1657 +) Add configuration option to build on Linux on both big-endian and
1659 [Ralf Baechle <ralf@uni-koblenz.de>]
1661 +) Give the OpenSSL applications more possibilities to make use of
1662 keys (public as well as private) handled by engines.
1665 +) Add OCSP code that comes from CertCo.
1668 +) Add VMS support for the Rijndael code.
1671 +) Added untested support for Nuron crypto accelerator.
1674 +) Add support for external cryptographic devices. This code was
1675 previously distributed separately as the "engine" branch.
1676 [Geoff Thorpe, Richard Levitte]
1678 +) Rework the filename-translation in the DSO code. It is now possible to
1679 have far greater control over how a "name" is turned into a filename
1680 depending on the operating environment and any oddities about the
1681 different shared library filenames on each system.
1684 +) Support threads on FreeBSD-elf in Configure.
1687 +) Add the possibility to create shared libraries on HP-UX
1690 +) Fix for SHA1 assembly problem with MASM: it produces
1691 warnings about corrupt line number information when assembling
1692 with debugging information. This is caused by the overlapping
1694 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
1697 NCONF_get_number() has no error checking at all. As a replacement,
1698 NCONF_get_number_e() is defined (_e for "error checking") and is
1699 promoted strongly. The old NCONF_get_number is kept around for
1700 binary backward compatibility.
1701 Make it possible for methods to load from something other than a BIO,
1702 by providing a function pointer that is given a name instead of a BIO.
1703 For example, this could be used to load configuration data from an
1707 *) Fix typo in get_cert_by_subject() in by_dir.c
1708 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
1710 *) Rework the system to generate shared libraries:
1712 - Make note of the expected extension for the shared libraries and
1713 if there is a need for symbolic links from for example libcrypto.so.0
1714 to libcrypto.so.0.9.7. There is extended info in Configure for
1717 - Make as few rebuilds of the shared libraries as possible.
1719 - Still avoid linking the OpenSSL programs with the shared libraries.
1721 - When installing, install the shared libraries separately from the
1725 +) Fix for non blocking accept BIOs. Added new I/O special reason
1726 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
1727 with non blocking I/O was not possible because no retry code was
1728 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
1732 +) Added the beginnings of Rijndael support.
1735 +) Fix for bug in DirectoryString mask setting. Add support for
1736 X509_NAME_print_ex() in 'req' and X509_print_ex() function
1737 to allow certificate printing to more controllable, additional
1738 'certopt' option to 'x509' to allow new printing options to be
1742 +) Clean old EAY MD5 hack from e_os.h.
1745 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
1747 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
1748 and not in SSL_clear because the latter is also used by the
1749 accept/connect functions; previously, the settings made by
1750 SSL_set_read_ahead would be lost during the handshake.
1751 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
1753 *) Correct util/mkdef.pl to be selective about disabled algorithms.
1754 Previously, it would create entries for disableed algorithms no
1758 *) Added several new manual pages for SSL_* function.
1761 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
1763 *) In ssl23_get_client_hello, generate an error message when faced
1764 with an initial SSL 3.0/TLS record that is too small to contain the
1765 first two bytes of the ClientHello message, i.e. client_version.
1766 (Note that this is a pathologic case that probably has never happened
1767 in real life.) The previous approach was to use the version number
1768 from the record header as a substitute; but our protocol choice
1769 should not depend on that one because it is not authenticated
1770 by the Finished messages.
1773 *) More robust randomness gathering functions for Windows.
1774 [Jeffrey Altman <jaltman@columbia.edu>]
1776 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
1777 not set then we don't setup the error code for issuer check errors
1778 to avoid possibly overwriting other errors which the callback does
1779 handle. If an application does set the flag then we assume it knows
1780 what it is doing and can handle the new informational codes
1784 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
1785 a general "ANY" type, as such it should be able to decode anything
1786 including tagged types. However it didn't check the class so it would
1787 wrongly interpret tagged types in the same way as their universal
1788 counterpart and unknown types were just rejected. Changed so that the
1789 tagged and unknown types are handled in the same way as a SEQUENCE:
1790 that is the encoding is stored intact. There is also a new type
1791 "V_ASN1_OTHER" which is used when the class is not universal, in this
1792 case we have no idea what the actual type is so we just lump them all
1796 *) On VMS, stdout may very well lead to a file that is written to
1797 in a record-oriented fashion. That means that every write() will
1798 write a separate record, which will be read separately by the
1799 programs trying to read from it. This can be very confusing.
1801 The solution is to put a BIO filter in the way that will buffer
1802 text until a linefeed is reached, and then write everything a
1803 line at a time, so every record written will be an actual line,
1804 not chunks of lines and not (usually doesn't happen, but I've
1805 seen it once) several lines in one record. BIO_f_linebuffer() is
1808 Currently, it's a VMS-only method, because that's where it has
1809 been tested well enough.
1812 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
1813 it can return incorrect results.
1814 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
1815 but it was in 0.9.6-beta[12].)
1818 *) Disable the check for content being present when verifying detached
1819 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
1820 include zero length content when signing messages.
1823 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
1824 BIO_ctrl (for BIO pairs).
1827 *) Add DSO method for VMS.
1830 *) Bug fix: Montgomery multiplication could produce results with the
1834 *) Add RPM specification openssl.spec and modify it to build three
1835 packages. The default package contains applications, application
1836 documentation and run-time libraries. The devel package contains
1837 include files, static libraries and function documentation. The
1838 doc package contains the contents of the doc directory. The original
1839 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
1842 *) Add a large number of documentation files for many SSL routines.
1843 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
1845 *) Add a configuration entry for Sony News 4.
1846 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
1848 *) Don't set the two most significant bits to one when generating a
1849 random number < q in the DSA library.
1852 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
1853 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
1854 the underlying transport is blocking) if a handshake took place.
1855 (The default behaviour is needed by applications such as s_client
1856 and s_server that use select() to determine when to use SSL_read;
1857 but for applications that know in advance when to expect data, it
1858 just makes things more complicated.)
1861 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
1865 *) Add a few more EBCDIC conditionals that make `req' and `x509'
1866 work better on such systems.
1867 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
1869 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
1870 Update PKCS12_parse() so it copies the friendlyName and the
1871 keyid to the certificates aux info.
1874 *) Fix bug in PKCS7_verify() which caused an infinite loop
1875 if there was more than one signature.
1876 [Sven Uszpelkat <su@celocom.de>]
1878 *) Major change in util/mkdef.pl to include extra information
1879 about each symbol, as well as presentig variables as well
1880 as functions. This change means that there's n more need
1881 to rebuild the .num files when some algorithms are excluded.
1884 *) Allow the verify time to be set by an application,
1885 rather than always using the current time.
1888 *) Phase 2 verify code reorganisation. The certificate
1889 verify code now looks up an issuer certificate by a
1890 number of criteria: subject name, authority key id
1891 and key usage. It also verifies self signed certificates
1892 by the same criteria. The main comparison function is
1893 X509_check_issued() which performs these checks.
1895 Lot of changes were necessary in order to support this
1896 without completely rewriting the lookup code.
1898 Authority and subject key identifier are now cached.
1900 The LHASH 'certs' is X509_STORE has now been replaced
1901 by a STACK_OF(X509_OBJECT). This is mainly because an
1902 LHASH can't store or retrieve multiple objects with
1903 the same hash value.
1905 As a result various functions (which were all internal
1906 use only) have changed to handle the new X509_STORE
1907 structure. This will break anything that messed round
1908 with X509_STORE internally.
1910 The functions X509_STORE_add_cert() now checks for an
1911 exact match, rather than just subject name.
1913 The X509_STORE API doesn't directly support the retrieval
1914 of multiple certificates matching a given criteria, however
1915 this can be worked round by performing a lookup first
1916 (which will fill the cache with candidate certificates)
1917 and then examining the cache for matches. This is probably
1918 the best we can do without throwing out X509_LOOKUP
1919 entirely (maybe later...).
1921 The X509_VERIFY_CTX structure has been enhanced considerably.
1923 All certificate lookup operations now go via a get_issuer()
1924 callback. Although this currently uses an X509_STORE it
1925 can be replaced by custom lookups. This is a simple way
1926 to bypass the X509_STORE hackery necessary to make this
1927 work and makes it possible to use more efficient techniques
1928 in future. A very simple version which uses a simple
1929 STACK for its trusted certificate store is also provided
1930 using X509_STORE_CTX_trusted_stack().
1932 The verify_cb() and verify() callbacks now have equivalents
1933 in the X509_STORE_CTX structure.
1935 X509_STORE_CTX also has a 'flags' field which can be used
1936 to customise the verify behaviour.
1939 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
1940 excludes S/MIME capabilities.
1943 *) When a certificate request is read in keep a copy of the
1944 original encoding of the signed data and use it when outputing
1945 again. Signatures then use the original encoding rather than
1946 a decoded, encoded version which may cause problems if the
1947 request is improperly encoded.
1950 *) For consistency with other BIO_puts implementations, call
1951 buffer_write(b, ...) directly in buffer_puts instead of calling
1954 In BIO_puts, increment b->num_write as in BIO_write.
1955 [Peter.Sylvester@EdelWeb.fr]
1957 *) Fix BN_mul_word for the case where the word is 0. (We have to use
1958 BN_zero, we may not return a BIGNUM with an array consisting of
1962 *) Avoid calling abort() from within the library when problems are
1963 detected, except if preprocessor symbols have been defined
1964 (such as REF_CHECK, BN_DEBUG etc.).
1967 *) New openssl application 'rsautl'. This utility can be
1968 used for low level RSA operations. DER public key
1969 BIO/fp routines also added.
1972 *) New Configure entry and patches for compiling on QNX 4.
1973 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
1975 *) A demo state-machine implementation was sponsored by
1976 Nuron (http://www.nuron.com/) and is now available in
1977 demos/state_machine.
1980 *) New options added to the 'dgst' utility for signature
1981 generation and verification.
1984 *) Unrecognized PKCS#7 content types are now handled via a
1985 catch all ASN1_TYPE structure. This allows unsupported
1986 types to be stored as a "blob" and an application can
1987 encode and decode it manually.
1990 *) Fix various signed/unsigned issues to make a_strex.c
1992 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
1994 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
1995 length if passed a buffer. ASN1_INTEGER_to_BN failed
1996 if passed a NULL BN and its argument was negative.
1997 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
1999 *) Modification to PKCS#7 encoding routines to output definite
2000 length encoding. Since currently the whole structures are in
2001 memory there's not real point in using indefinite length
2002 constructed encoding. However if OpenSSL is compiled with
2003 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
2006 *) Added BIO_vprintf() and BIO_vsnprintf().
2009 *) Added more prefixes to parse for in the the strings written
2010 through a logging bio, to cover all the levels that are available
2011 through syslog. The prefixes are now:
2013 PANIC, EMERG, EMR => LOG_EMERG
2014 ALERT, ALR => LOG_ALERT
2015 CRIT, CRI => LOG_CRIT
2016 ERROR, ERR => LOG_ERR
2017 WARNING, WARN, WAR => LOG_WARNING
2018 NOTICE, NOTE, NOT => LOG_NOTICE
2019 INFO, INF => LOG_INFO
2020 DEBUG, DBG => LOG_DEBUG
2022 and as before, if none of those prefixes are present at the
2023 beginning of the string, LOG_ERR is chosen.
2025 On Win32, the LOG_* levels are mapped according to this:
2027 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
2028 LOG_WARNING => EVENTLOG_WARNING_TYPE
2029 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
2033 *) Made it possible to reconfigure with just the configuration
2034 argument "reconf" or "reconfigure". The command line arguments
2035 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
2036 and are retrieved from there when reconfiguring.
2040 [Assar Westerlund <assar@sics.se>, Richard Levitte]
2042 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
2045 *) The obj_dat.pl script was messing up the sorting of object
2046 names. The reason was that it compared the quoted version
2047 of strings as a result "OCSP" > "OCSP Signing" because
2048 " > SPACE. Changed script to store unquoted versions of
2049 names and add quotes on output. It was also omitting some
2050 names from the lookup table if they were given a default
2051 value (that is if SN is missing it is given the same
2052 value as LN and vice versa), these are now added on the
2053 grounds that if an object has a name we should be able to
2054 look it up. Finally added warning output when duplicate
2055 short or long names are found.
2058 *) Changes needed for Tandem NSK.
2059 [Scott Uroff <scott@xypro.com>]
2061 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
2062 RSA_padding_check_SSLv23(), special padding was never detected
2063 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
2064 version rollback attacks was not effective.
2066 In s23_clnt.c, don't use special rollback-attack detection padding
2067 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
2068 client; similarly, in s23_srvr.c, don't do the rollback check if
2069 SSL 2.0 is the only protocol enabled in the server.
2072 *) Make it possible to get hexdumps of unprintable data with 'openssl
2073 asn1parse'. By implication, the functions ASN1_parse_dump() and
2074 BIO_dump_indent() are added.
2077 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
2078 these print out strings and name structures based on various
2079 flags including RFC2253 support and proper handling of
2080 multibyte characters. Added options to the 'x509' utility
2081 to allow the various flags to be set.
2084 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
2085 Also change the functions X509_cmp_current_time() and
2086 X509_gmtime_adj() work with an ASN1_TIME structure,
2087 this will enable certificates using GeneralizedTime in validity
2088 dates to be checked.
2091 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
2092 negative public key encodings) on by default,
2093 NO_NEG_PUBKEY_BUG can be set to disable it.
2096 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
2097 content octets. An i2c_ASN1_OBJECT is unnecessary because
2098 the encoding can be trivially obtained from the structure.
2101 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
2102 not read locks (CRYPTO_r_[un]lock).
2105 *) A first attempt at creating official support for shared
2106 libraries through configuration. I've kept it so the
2107 default is static libraries only, and the OpenSSL programs
2108 are always statically linked for now, but there are
2109 preparations for dynamic linking in place.
2110 This has been tested on Linux and Tru64.
2113 *) Randomness polling function for Win9x, as described in:
2114 Peter Gutmann, Software Generation of Practically Strong
2118 *) Fix so PRNG is seeded in req if using an already existing
2122 *) New options to smime application. -inform and -outform
2123 allow alternative formats for the S/MIME message including
2124 PEM and DER. The -content option allows the content to be
2125 specified separately. This should allow things like Netscape
2126 form signing output easier to verify.
2129 *) Fix the ASN1 encoding of tags using the 'long form'.
2132 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
2133 STRING types. These convert content octets to and from the
2134 underlying type. The actual tag and length octets are
2135 already assumed to have been read in and checked. These
2136 are needed because all other string types have virtually
2137 identical handling apart from the tag. By having versions
2138 of the ASN1 functions that just operate on content octets
2139 IMPLICIT tagging can be handled properly. It also allows
2140 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
2141 and ASN1_INTEGER are identical apart from the tag.
2144 *) Change the handling of OID objects as follows:
2146 - New object identifiers are inserted in objects.txt, following
2147 the syntax given in objects.README.
2148 - objects.pl is used to process obj_mac.num and create a new
2150 - obj_dat.pl is used to create a new obj_dat.h, using the data in
2153 This is currently kind of a hack, and the perl code in objects.pl
2154 isn't very elegant, but it works as I intended. The simplest way
2155 to check that it worked correctly is to look in obj_dat.h and
2156 check the array nid_objs and make sure the objects haven't moved
2157 around (this is important!). Additions are OK, as well as
2158 consistent name changes.
2161 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
2164 *) Addition of the command line parameter '-rand file' to 'openssl req'.
2165 The given file adds to whatever has already been seeded into the
2166 random pool through the RANDFILE configuration file option or
2167 environment variable, or the default random state file.
2170 *) mkstack.pl now sorts each macro group into lexical order.
2171 Previously the output order depended on the order the files
2172 appeared in the directory, resulting in needless rewriting
2176 *) Patches to make OpenSSL compile under Win32 again. Mostly
2177 work arounds for the VC++ problem that it treats func() as
2178 func(void). Also stripped out the parts of mkdef.pl that
2179 added extra typesafe functions: these no longer exist.
2182 *) Reorganisation of the stack code. The macros are now all
2183 collected in safestack.h . Each macro is defined in terms of
2184 a "stack macro" of the form SKM_<name>(type, a, b). The
2185 DEBUG_SAFESTACK is now handled in terms of function casts,
2186 this has the advantage of retaining type safety without the
2187 use of additional functions. If DEBUG_SAFESTACK is not defined
2188 then the non typesafe macros are used instead. Also modified the
2189 mkstack.pl script to handle the new form. Needs testing to see
2190 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
2191 the default if no major problems. Similar behaviour for ASN1_SET_OF
2192 and PKCS12_STACK_OF.
2195 *) When some versions of IIS use the 'NET' form of private key the
2196 key derivation algorithm is different. Normally MD5(password) is
2197 used as a 128 bit RC4 key. In the modified case
2198 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
2199 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
2200 as the old Netscape_RSA functions except they have an additional
2201 'sgckey' parameter which uses the modified algorithm. Also added
2202 an -sgckey command line option to the rsa utility. Thanks to
2203 Adrian Peck <bertie@ncipher.com> for posting details of the modified
2204 algorithm to openssl-dev.
2207 *) The evp_local.h macros were using 'c.##kname' which resulted in
2208 invalid expansion on some systems (SCO 5.0.5 for example).
2209 Corrected to 'c.kname'.
2210 [Phillip Porch <root@theporch.com>]
2212 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
2213 a STACK of email addresses from a certificate or request, these look
2214 in the subject name and the subject alternative name extensions and
2215 omit any duplicate addresses.
2218 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
2219 This makes DSA verification about 2 % faster.
2222 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
2223 (meaning that now 2^5 values will be precomputed, which is only 4 KB
2224 plus overhead for 1024 bit moduli).
2225 This makes exponentiations about 0.5 % faster for 1024 bit
2226 exponents (as measured by "openssl speed rsa2048").
2229 *) Rename memory handling macros to avoid conflicts with other
2231 Malloc => OPENSSL_malloc
2232 Malloc_locked => OPENSSL_malloc_locked
2233 Realloc => OPENSSL_realloc
2234 Free => OPENSSL_free
2237 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
2238 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
2241 *) CygWin32 support.
2242 [John Jarvie <jjarvie@newsguy.com>]
2244 *) The type-safe stack code has been rejigged. It is now only compiled
2245 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
2246 by default all type-specific stack functions are "#define"d back to
2247 standard stack functions. This results in more streamlined output
2248 but retains the type-safety checking possibilities of the original
2252 *) The STACK code has been cleaned up, and certain type declarations
2253 that didn't make a lot of sense have been brought in line. This has
2254 also involved a cleanup of sorts in safestack.h to more correctly
2255 map type-safe stack functions onto their plain stack counterparts.
2256 This work has also resulted in a variety of "const"ifications of
2257 lots of the code, especially "_cmp" operations which should normally
2258 be prototyped with "const" parameters anyway.
2261 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
2262 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
2263 (The PRNG state consists of two parts, the large pool 'state' and 'md',
2264 where all of 'md' is used each time the PRNG is used, but 'state'
2265 is used only indexed by a cyclic counter. As entropy may not be
2266 well distributed from the beginning, 'md' is important as a
2267 chaining variable. However, the output function chains only half
2268 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
2269 all of 'md', and seeding with STATE_SIZE dummy bytes will result
2270 in all of 'state' being rewritten, with the new values depending
2271 on virtually all of 'md'. This overcomes the 80 bit limitation.)
2274 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
2275 the handshake is continued after ssl_verify_cert_chain();
2276 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
2277 can lead to 'unexplainable' connection aborts later.
2278 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
2280 *) Major EVP API cipher revision.
2281 Add hooks for extra EVP features. This allows various cipher
2282 parameters to be set in the EVP interface. Support added for variable
2283 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
2284 setting of RC2 and RC5 parameters.
2286 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
2289 Remove lots of duplicated code from the EVP library. For example *every*
2290 cipher init() function handles the 'iv' in the same way according to the
2291 cipher mode. They also all do nothing if the 'key' parameter is NULL and
2292 for CFB and OFB modes they zero ctx->num.
2294 New functionality allows removal of S/MIME code RC2 hack.
2296 Most of the routines have the same form and so can be declared in terms
2299 By shifting this to the top level EVP_CipherInit() it can be removed from
2300 all individual ciphers. If the cipher wants to handle IVs or keys
2301 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
2304 Change lots of functions like EVP_EncryptUpdate() to now return a
2305 value: although software versions of the algorithms cannot fail
2306 any installed hardware versions can.
2309 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
2310 this option is set, tolerate broken clients that send the negotiated
2311 protocol version number instead of the requested protocol version
2315 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
2316 i.e. non-zero for export ciphersuites, zero otherwise.
2317 Previous versions had this flag inverted, inconsistent with
2318 rsa_tmp_cb (..._TMP_RSA_CB).
2319 [Bodo Moeller; problem reported by Amit Chopra]
2321 *) Add missing DSA library text string. Work around for some IIS
2322 key files with invalid SEQUENCE encoding.
2325 *) Add a document (doc/standards.txt) that list all kinds of standards
2326 and so on that are implemented in OpenSSL.
2329 *) Enhance c_rehash script. Old version would mishandle certificates
2330 with the same subject name hash and wouldn't handle CRLs at all.
2331 Added -fingerprint option to crl utility, to support new c_rehash
2335 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
2338 *) Fix for SSL server purpose checking. Server checking was
2339 rejecting certificates which had extended key usage present
2340 but no ssl client purpose.
2341 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
2343 *) Make PKCS#12 code work with no password. The PKCS#12 spec
2344 is a little unclear about how a blank password is handled.
2345 Since the password in encoded as a BMPString with terminating
2346 double NULL a zero length password would end up as just the
2347 double NULL. However no password at all is different and is
2348 handled differently in the PKCS#12 key generation code. NS
2349 treats a blank password as zero length. MSIE treats it as no
2350 password on export: but it will try both on import. We now do
2351 the same: PKCS12_parse() tries zero length and no password if
2352 the password is set to "" or NULL (NULL is now a valid password:
2353 it wasn't before) as does the pkcs12 application.
2356 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
2357 perror when PEM_read_bio_X509_REQ fails, the error message must
2358 be obtained from the error queue.
2361 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
2362 it in ERR_remove_state if appropriate, and change ERR_get_state
2363 accordingly to avoid race conditions (this is necessary because
2364 thread_hash is no longer constant once set).
2367 *) Bugfix for linux-elf makefile.one.
2370 *) RSA_get_default_method() will now cause a default
2371 RSA_METHOD to be chosen if one doesn't exist already.
2372 Previously this was only set during a call to RSA_new()
2373 or RSA_new_method(NULL) meaning it was possible for
2374 RSA_get_default_method() to return NULL.
2377 *) Added native name translation to the existing DSO code
2378 that will convert (if the flag to do so is set) filenames
2379 that are sufficiently small and have no path information
2380 into a canonical native form. Eg. "blah" converted to
2381 "libblah.so" or "blah.dll" etc.
2384 *) New function ERR_error_string_n(e, buf, len) which is like
2385 ERR_error_string(e, buf), but writes at most 'len' bytes
2386 including the 0 terminator. For ERR_error_string_n, 'buf'
2388 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
2390 *) CONF library reworked to become more general. A new CONF
2391 configuration file reader "class" is implemented as well as a
2392 new functions (NCONF_*, for "New CONF") to handle it. The now
2393 old CONF_* functions are still there, but are reimplemented to
2394 work in terms of the new functions. Also, a set of functions
2395 to handle the internal storage of the configuration data is
2396 provided to make it easier to write new configuration file
2397 reader "classes" (I can definitely see something reading a
2398 configuration file in XML format, for example), called _CONF_*,
2399 or "the configuration storage API"...
2401 The new configuration file reading functions are:
2403 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
2404 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
2406 NCONF_default, NCONF_WIN32
2408 NCONF_dump_fp, NCONF_dump_bio
2410 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
2411 NCONF_new creates a new CONF object. This works in the same way
2412 as other interfaces in OpenSSL, like the BIO interface.
2413 NCONF_dump_* dump the internal storage of the configuration file,
2414 which is useful for debugging. All other functions take the same
2415 arguments as the old CONF_* functions wth the exception of the
2416 first that must be a `CONF *' instead of a `LHASH *'.
2418 To make it easer to use the new classes with the old CONF_* functions,
2419 the function CONF_set_default_method is provided.
2422 *) Add '-tls1' option to 'openssl ciphers', which was already
2423 mentioned in the documentation but had not been implemented.
2424 (This option is not yet really useful because even the additional
2425 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
2428 *) Initial DSO code added into libcrypto for letting OpenSSL (and
2429 OpenSSL-based applications) load shared libraries and bind to
2430 them in a portable way.
2431 [Geoff Thorpe, with contributions from Richard Levitte]
2433 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
2435 *) Make sure _lrotl and _lrotr are only used with MSVC.
2437 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
2438 (the default implementation of RAND_status).
2440 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
2441 to '-clrext' (= clear extensions), as intended and documented.
2442 [Bodo Moeller; inconsistency pointed out by Michael Attili
2445 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
2446 was larger than the MD block size.
2447 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
2449 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
2450 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
2451 using the passed key: if the passed key was a private key the result
2452 of X509_print(), for example, would be to print out all the private key
2456 *) des_quad_cksum() byte order bug fix.
2457 [Ulf Möller, using the problem description in krb4-0.9.7, where
2458 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
2460 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
2462 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
2464 *) For easily testing in shell scripts whether some command
2465 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
2466 returns with exit code 0 iff no command of the given name is available.
2467 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
2468 the output goes to stdout and nothing is printed to stderr.
2469 Additional arguments are always ignored.
2471 Since for each cipher there is a command of the same name,
2472 the 'no-cipher' compilation switches can be tested this way.
2474 ('openssl no-XXX' is not able to detect pseudo-commands such
2475 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
2478 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
2481 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
2482 is set; it will be thrown away anyway because each handshake creates
2484 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
2485 to parameters -- in previous versions (since OpenSSL 0.9.3) the
2486 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
2487 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
2490 *) New s_client option -ign_eof: EOF at stdin is ignored, and
2491 'Q' and 'R' lose their special meanings (quit/renegotiate).
2492 This is part of what -quiet does; unlike -quiet, -ign_eof
2493 does not suppress any output.
2496 *) Add compatibility options to the purpose and trust code. The
2497 purpose X509_PURPOSE_ANY is "any purpose" which automatically
2498 accepts a certificate or CA, this was the previous behaviour,
2499 with all the associated security issues.
2501 X509_TRUST_COMPAT is the old trust behaviour: only and
2502 automatically trust self signed roots in certificate store. A
2503 new trust setting X509_TRUST_DEFAULT is used to specify that
2504 a purpose has no associated trust setting and it should instead
2505 use the value in the default purpose.
2508 *) Fix the PKCS#8 DSA private key code so it decodes keys again
2509 and fix a memory leak.
2512 *) In util/mkerr.pl (which implements 'make errors'), preserve
2513 reason strings from the previous version of the .c file, as
2514 the default to have only downcase letters (and digits) in
2515 automatically generated reasons codes is not always appropriate.
2518 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
2519 using strerror. Previously, ERR_reason_error_string() returned
2520 library names as reason strings for SYSerr; but SYSerr is a special
2521 case where small numbers are errno values, not library numbers.
2524 *) Add '-dsaparam' option to 'openssl dhparam' application. This
2525 converts DSA parameters into DH parameters. (When creating parameters,
2526 DSA_generate_parameters is used.)
2529 *) Include 'length' (recommended exponent length) in C code generated
2530 by 'openssl dhparam -C'.
2533 *) The second argument to set_label in perlasm was already being used
2534 so couldn't be used as a "file scope" flag. Moved to third argument
2538 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
2539 instead of RAND_bytes for encryption IVs and salts.
2542 *) Include RAND_status() into RAND_METHOD instead of implementing
2543 it only for md_rand.c Otherwise replacing the PRNG by calling
2544 RAND_set_rand_method would be impossible.
2547 *) Don't let DSA_generate_key() enter an infinite loop if the random
2548 number generation fails.
2551 *) New 'rand' application for creating pseudo-random output.
2554 *) Added configuration support for Linux/IA64
2555 [Rolf Haberrecker <rolf@suse.de>]
2557 *) Assembler module support for Mingw32.
2560 *) Shared library support for HPUX (in shlib/).
2561 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
2563 *) Shared library support for Solaris gcc.
2564 [Lutz Behnke <behnke@trustcenter.de>]
2566 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
2568 *) PKCS7_encrypt() was adding text MIME headers twice because they
2569 were added manually and by SMIME_crlf_copy().
2572 *) In bntest.c don't call BN_rand with zero bits argument.
2573 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
2575 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
2576 case was implemented. This caused BN_div_recp() to fail occasionally.
2579 *) Add an optional second argument to the set_label() in the perl
2580 assembly language builder. If this argument exists and is set
2581 to 1 it signals that the assembler should use a symbol whose
2582 scope is the entire file, not just the current function. This
2583 is needed with MASM which uses the format label:: for this scope.
2584 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
2586 *) Change the ASN1 types so they are typedefs by default. Before
2587 almost all types were #define'd to ASN1_STRING which was causing
2588 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
2592 *) Change names of new functions to the new get1/get0 naming
2593 convention: After 'get1', the caller owns a reference count
2594 and has to call ..._free; 'get0' returns a pointer to some
2595 data structure without incrementing reference counters.
2596 (Some of the existing 'get' functions increment a reference
2597 counter, some don't.)
2598 Similarly, 'set1' and 'add1' functions increase reference
2599 counters or duplicate objects.
2602 *) Allow for the possibility of temp RSA key generation failure:
2603 the code used to assume it always worked and crashed on failure.
2606 *) Fix potential buffer overrun problem in BIO_printf().
2607 [Ulf Möller, using public domain code by Patrick Powell; problem
2608 pointed out by David Sacerdote <das33@cornell.edu>]
2610 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
2611 RAND_egd() and RAND_status(). In the command line application,
2612 the EGD socket can be specified like a seed file using RANDFILE
2616 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
2617 Some CAs (e.g. Verisign) distribute certificates in this form.
2620 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
2621 list to exclude them. This means that no special compilation option
2622 is needed to use anonymous DH: it just needs to be included in the
2626 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
2627 EVP_MD_type. The old functionality is available in a new macro called
2628 EVP_MD_md(). Change code that uses it and update docs.
2631 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
2632 where the 'void *' argument is replaced by a function pointer argument.
2633 Previously 'void *' was abused to point to functions, which works on
2634 many platforms, but is not correct. As these functions are usually
2635 called by macros defined in OpenSSL header files, most source code
2636 should work without changes.
2639 *) <openssl/opensslconf.h> (which is created by Configure) now contains
2640 sections with information on -D... compiler switches used for
2641 compiling the library so that applications can see them. To enable
2642 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
2643 must be defined. E.g.,
2644 #define OPENSSL_ALGORITHM_DEFINES
2645 #include <openssl/opensslconf.h>
2646 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
2647 [Richard Levitte, Ulf and Bodo Möller]
2649 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
2653 *) Change the 'other' type in certificate aux info to a STACK_OF
2654 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
2655 the required ASN1 format: arbitrary types determined by an OID.
2658 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
2659 argument to 'req'. This is not because the function is newer or
2660 better than others it just uses the work 'NEW' in the certificate
2661 request header lines. Some software needs this.
2664 *) Reorganise password command line arguments: now passwords can be
2665 obtained from various sources. Delete the PEM_cb function and make
2666 it the default behaviour: i.e. if the callback is NULL and the
2667 usrdata argument is not NULL interpret it as a null terminated pass
2668 phrase. If usrdata and the callback are NULL then the pass phrase
2669 is prompted for as usual.
2672 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
2673 the support is automatically enabled. The resulting binaries will
2674 autodetect the card and use it if present.
2675 [Ben Laurie and Compaq Inc.]
2677 *) Work around for Netscape hang bug. This sends certificate request
2678 and server done in one record. Since this is perfectly legal in the
2679 SSL/TLS protocol it isn't a "bug" option and is on by default. See
2680 the bugs/SSLv3 entry for more info.
2683 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
2686 *) Add -rand argument to smime and pkcs12 applications and read/write
2690 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
2693 *) Add command line password options to the remaining applications.
2696 *) Bug fix for BN_div_recp() for numerators with an even number of
2700 *) More tests in bntest.c, and changed test_bn output.
2703 *) ./config recognizes MacOS X now.
2706 *) Bug fix for BN_div() when the first words of num and divsor are
2707 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
2710 *) Add support for various broken PKCS#8 formats, and command line
2711 options to produce them.
2714 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
2715 get temporary BIGNUMs from a BN_CTX.
2718 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
2722 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
2723 include a #define from the old name to the new. The original intent
2724 was that statically linked binaries could for example just call
2725 SSLeay_add_all_ciphers() to just add ciphers to the table and not
2726 link with digests. This never worked becayse SSLeay_add_all_digests()
2727 and SSLeay_add_all_ciphers() were in the same source file so calling
2728 one would link with the other. They are now in separate source files.
2731 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
2734 *) Use a less unusual form of the Miller-Rabin primality test (it used
2735 a binary algorithm for exponentiation integrated into the Miller-Rabin
2736 loop, our standard modexp algorithms are faster).
2739 *) Support for the EBCDIC character set completed.
2740 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
2742 *) Source code cleanups: use const where appropriate, eliminate casts,
2743 use void * instead of char * in lhash.
2746 *) Bugfix: ssl3_send_server_key_exchange was not restartable
2747 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
2748 this the server could overwrite ephemeral keys that the client
2752 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
2753 using 50 iterations of the Rabin-Miller test.
2755 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
2756 iterations of the Rabin-Miller test as required by the appendix
2757 to FIPS PUB 186[-1]) instead of DSA_is_prime.
2758 As BN_is_prime_fasttest includes trial division, DSA parameter
2759 generation becomes much faster.
2761 This implies a change for the callback functions in DSA_is_prime
2762 and DSA_generate_parameters: The callback function is called once
2763 for each positive witness in the Rabin-Miller test, not just
2764 occasionally in the inner loop; and the parameters to the
2765 callback function now provide an iteration count for the outer
2766 loop rather than for the current invocation of the inner loop.
2767 DSA_generate_parameters additionally can call the callback
2768 function with an 'iteration count' of -1, meaning that a
2769 candidate has passed the trial division test (when q is generated
2770 from an application-provided seed, trial division is skipped).
2773 *) New function BN_is_prime_fasttest that optionally does trial
2774 division before starting the Rabin-Miller test and has
2775 an additional BN_CTX * argument (whereas BN_is_prime always
2776 has to allocate at least one BN_CTX).
2777 'callback(1, -1, cb_arg)' is called when a number has passed the
2778 trial division stage.
2781 *) Fix for bug in CRL encoding. The validity dates weren't being handled
2785 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
2788 *) New function BN_pseudo_rand().
2791 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
2792 bignum version of BN_from_montgomery() with the working code from
2793 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
2797 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
2798 made it impossible to use the same SSL_SESSION data structure in
2799 SSL2 clients in multiple threads.
2802 *) The return value of RAND_load_file() no longer counts bytes obtained
2803 by stat(). RAND_load_file(..., -1) is new and uses the complete file
2804 to seed the PRNG (previously an explicit byte count was required).
2805 [Ulf Möller, Bodo Möller]
2807 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
2808 used (char *) instead of (void *) and had casts all over the place.
2811 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
2814 *) Retain source code compatibility for BN_prime_checks macro:
2815 BN_is_prime(..., BN_prime_checks, ...) now uses
2816 BN_prime_checks_for_size to determine the appropriate number of
2817 Rabin-Miller iterations.
2820 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
2821 DH_CHECK_P_NOT_SAFE_PRIME.
2822 (Check if this is true? OpenPGP calls them "strong".)
2825 *) Merge the functionality of "dh" and "gendh" programs into a new program
2826 "dhparam". The old programs are retained for now but will handle DH keys
2827 (instead of parameters) in future.
2830 *) Make the ciphers, s_server and s_client programs check the return values
2831 when a new cipher list is set.
2834 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
2835 ciphers. Before when the 56bit ciphers were enabled the sorting was
2838 The syntax for the cipher sorting has been extended to support sorting by
2839 cipher-strength (using the strength_bits hard coded in the tables).
2840 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
2842 Fix a bug in the cipher-command parser: when supplying a cipher command
2843 string with an "undefined" symbol (neither command nor alphanumeric
2844 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
2845 an error is flagged.
2847 Due to the strength-sorting extension, the code of the
2848 ssl_create_cipher_list() function was completely rearranged. I hope that
2849 the readability was also increased :-)
2850 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
2852 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
2853 for the first serial number and places 2 in the serial number file. This
2854 avoids problems when the root CA is created with serial number zero and
2855 the first user certificate has the same issuer name and serial number
2859 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
2860 the new code. Add documentation for this stuff.
2863 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
2864 X509_*() to X509at_*() on the grounds that they don't handle X509
2865 structures and behave in an analagous way to the X509v3 functions:
2866 they shouldn't be called directly but wrapper functions should be used
2869 So we also now have some wrapper functions that call the X509at functions
2870 when passed certificate requests. (TO DO: similar things can be done with
2871 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
2872 things. Some of these need some d2i or i2d and print functionality
2873 because they handle more complex structures.)
2876 *) Add missing #ifndefs that caused missing symbols when building libssl
2877 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
2878 NO_RSA in ssl/s2*.c.
2879 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
2881 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
2882 has a return value which indicates the quality of the random data
2883 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
2884 error queue. New function RAND_pseudo_bytes() generates output that is
2885 guaranteed to be unique but not unpredictable. RAND_add is like
2886 RAND_seed, but takes an extra argument for an entropy estimate
2887 (RAND_seed always assumes full entropy).
2890 *) Do more iterations of Rabin-Miller probable prime test (specifically,
2891 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
2892 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
2893 in crypto/bn/bn_prime.c for the complete table). This guarantees a
2894 false-positive rate of at most 2^-80 for random input.
2897 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
2900 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
2901 in the 0.9.5 release), this returns the chain
2902 from an X509_CTX structure with a dup of the stack and all
2903 the X509 reference counts upped: so the stack will exist
2904 after X509_CTX_cleanup() has been called. Modify pkcs12.c
2907 Also make SSL_SESSION_print() print out the verify return
2911 *) Add manpage for the pkcs12 command. Also change the default
2912 behaviour so MAC iteration counts are used unless the new
2913 -nomaciter option is used. This improves file security and
2914 only older versions of MSIE (4.0 for example) need it.
2917 *) Honor the no-xxx Configure options when creating .DEF files.
2920 *) Add PKCS#10 attributes to field table: challengePassword,
2921 unstructuredName and unstructuredAddress. These are taken from
2922 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
2923 international characters are used.
2925 More changes to X509_ATTRIBUTE code: allow the setting of types
2926 based on strings. Remove the 'loc' parameter when adding
2927 attributes because these will be a SET OF encoding which is sorted
2931 *) Initial changes to the 'req' utility to allow request generation
2932 automation. This will allow an application to just generate a template
2933 file containing all the field values and have req construct the
2936 Initial support for X509_ATTRIBUTE handling. Stacks of these are
2937 used all over the place including certificate requests and PKCS#7
2938 structures. They are currently handled manually where necessary with
2939 some primitive wrappers for PKCS#7. The new functions behave in a
2940 manner analogous to the X509 extension functions: they allow
2941 attributes to be looked up by NID and added.
2943 Later something similar to the X509V3 code would be desirable to
2944 automatically handle the encoding, decoding and printing of the
2945 more complex types. The string types like challengePassword can
2946 be handled by the string table functions.
2948 Also modified the multi byte string table handling. Now there is
2949 a 'global mask' which masks out certain types. The table itself
2950 can use the flag STABLE_NO_MASK to ignore the mask setting: this
2951 is useful when for example there is only one permissible type
2952 (as in countryName) and using the mask might result in no valid
2956 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
2957 SSL_get_peer_finished to allow applications to obtain the latest
2958 Finished messages sent to the peer or expected from the peer,
2959 respectively. (SSL_get_peer_finished is usually the Finished message
2960 actually received from the peer, otherwise the protocol will be aborted.)
2962 As the Finished message are message digests of the complete handshake
2963 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
2964 be used for external authentication procedures when the authentication
2965 provided by SSL/TLS is not desired or is not enough.
2968 *) Enhanced support for Alpha Linux is added. Now ./config checks if
2969 the host supports BWX extension and if Compaq C is present on the
2970 $PATH. Just exploiting of the BWX extension results in 20-30%
2971 performance kick for some algorithms, e.g. DES and RC4 to mention
2972 a couple. Compaq C in turn generates ~20% faster code for MD5 and
2976 *) Add support for MS "fast SGC". This is arguably a violation of the
2977 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
2978 weak crypto and after checking the certificate is SGC a second one
2979 with strong crypto. MS SGC stops the first handshake after receiving
2980 the server certificate message and sends a second client hello. Since
2981 a server will typically do all the time consuming operations before
2982 expecting any further messages from the client (server key exchange
2983 is the most expensive) there is little difference between the two.
2985 To get OpenSSL to support MS SGC we have to permit a second client
2986 hello message after we have sent server done. In addition we have to
2987 reset the MAC if we do get this second client hello.
2990 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
2991 if a DER encoded private key is RSA or DSA traditional format. Changed
2992 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
2993 format DER encoded private key. Newer code should use PKCS#8 format which
2994 has the key type encoded in the ASN1 structure. Added DER private key
2995 support to pkcs8 application.
2998 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
2999 ciphersuites has been selected (as required by the SSL 3/TLS 1
3000 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
3001 is set, we interpret this as a request to violate the specification
3002 (the worst that can happen is a handshake failure, and 'correct'
3003 behaviour would result in a handshake failure anyway).
3006 *) In SSL_CTX_add_session, take into account that there might be multiple
3007 SSL_SESSION structures with the same session ID (e.g. when two threads
3008 concurrently obtain them from an external cache).
3009 The internal cache can handle only one SSL_SESSION with a given ID,
3010 so if there's a conflict, we now throw out the old one to achieve
3014 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
3015 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
3016 some routines that use cipher OIDs: some ciphers do not have OIDs
3017 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
3021 *) Simplify the trust setting structure and code. Now we just have
3022 two sequences of OIDs for trusted and rejected settings. These will
3023 typically have values the same as the extended key usage extension
3024 and any application specific purposes.
3026 The trust checking code now has a default behaviour: it will just
3027 check for an object with the same NID as the passed id. Functions can
3028 be provided to override either the default behaviour or the behaviour
3029 for a given id. SSL client, server and email already have functions
3030 in place for compatibility: they check the NID and also return "trusted"
3031 if the certificate is self signed.
3034 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
3035 traditional format into an EVP_PKEY structure.
3038 *) Add a password callback function PEM_cb() which either prompts for
3039 a password if usr_data is NULL or otherwise assumes it is a null
3040 terminated password. Allow passwords to be passed on command line
3041 environment or config files in a few more utilities.
3044 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
3045 keys. Add some short names for PKCS#8 PBE algorithms and allow them
3046 to be specified on the command line for the pkcs8 and pkcs12 utilities.
3047 Update documentation.
3050 *) Support for ASN1 "NULL" type. This could be handled before by using
3051 ASN1_TYPE but there wasn't any function that would try to read a NULL
3052 and produce an error if it couldn't. For compatibility we also have
3053 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
3054 don't allocate anything because they don't need to.
3057 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
3059 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
3061 *) Rebuild of the memory allocation routines used by OpenSSL code and
3062 possibly others as well. The purpose is to make an interface that
3063 provide hooks so anyone can build a separate set of allocation and
3064 deallocation routines to be used by OpenSSL, for example memory
3065 pool implementations, or something else, which was previously hard
3066 since Malloc(), Realloc() and Free() were defined as macros having
3067 the values malloc, realloc and free, respectively (except for Win32
3068 compilations). The same is provided for memory debugging code.
3069 OpenSSL already comes with functionality to find memory leaks, but
3070 this gives people a chance to debug other memory problems.
3072 With these changes, a new set of functions and macros have appeared:
3074 CRYPTO_set_mem_debug_functions() [F]
3075 CRYPTO_get_mem_debug_functions() [F]
3076 CRYPTO_dbg_set_options() [F]
3077 CRYPTO_dbg_get_options() [F]
3078 CRYPTO_malloc_debug_init() [M]
3080 The memory debug functions are NULL by default, unless the library
3081 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
3082 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
3083 gives the standard debugging functions that come with OpenSSL) or
3084 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
3085 provided by the library user) must be used. When the standard
3086 debugging functions are used, CRYPTO_dbg_set_options can be used to
3087 request additional information:
3088 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
3089 the CRYPTO_MDEBUG_xxx macro when compiling the library.
3091 Also, things like CRYPTO_set_mem_functions will always give the
3092 expected result (the new set of functions is used for allocation
3093 and deallocation) at all times, regardless of platform and compiler
3096 To finish it up, some functions that were never use in any other
3097 way than through macros have a new API and new semantic:
3100 CRYPTO_dbg_realloc()
3103 All macros of value have retained their old syntax.
3104 [Richard Levitte and Bodo Moeller]
3106 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
3107 ordering of SMIMECapabilities wasn't in "strength order" and there
3108 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
3112 *) Some ASN1 types with illegal zero length encoding (INTEGER,
3113 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
3114 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
3116 *) Merge in my S/MIME library for OpenSSL. This provides a simple
3117 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
3118 functionality to handle multipart/signed properly) and a utility
3119 called 'smime' to call all this stuff. This is based on code I
3120 originally wrote for Celo who have kindly allowed it to be
3121 included in OpenSSL.
3124 *) Add variants des_set_key_checked and des_set_key_unchecked of
3125 des_set_key (aka des_key_sched). Global variable des_check_key
3126 decides which of these is called by des_set_key; this way
3127 des_check_key behaves as it always did, but applications and
3128 the library itself, which was buggy for des_check_key == 1,
3129 have a cleaner way to pick the version they need.
3132 *) New function PKCS12_newpass() which changes the password of a
3136 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
3137 dynamic mix. In both cases the ids can be used as an index into the
3138 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
3139 functions so they accept a list of the field values and the
3140 application doesn't need to directly manipulate the X509_TRUST
3144 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
3148 *) Modify the way the V3 extension code looks up extensions. This now
3149 works in a similar way to the object code: we have some "standard"
3150 extensions in a static table which is searched with OBJ_bsearch()
3151 and the application can add dynamic ones if needed. The file
3152 crypto/x509v3/ext_dat.h now has the info: this file needs to be
3153 updated whenever a new extension is added to the core code and kept
3154 in ext_nid order. There is a simple program 'tabtest.c' which checks
3155 this. New extensions are not added too often so this file can readily
3156 be maintained manually.
3158 There are two big advantages in doing things this way. The extensions
3159 can be looked up immediately and no longer need to be "added" using
3160 X509V3_add_standard_extensions(): this function now does nothing.
3161 [Side note: I get *lots* of email saying the extension code doesn't
3162 work because people forget to call this function]
3163 Also no dynamic allocation is done unless new extensions are added:
3164 so if we don't add custom extensions there is no need to call
3165 X509V3_EXT_cleanup().
3168 *) Modify enc utility's salting as follows: make salting the default. Add a
3169 magic header, so unsalted files fail gracefully instead of just decrypting
3170 to garbage. This is because not salting is a big security hole, so people
3171 should be discouraged from doing it.
3174 *) Fixes and enhancements to the 'x509' utility. It allowed a message
3175 digest to be passed on the command line but it only used this
3176 parameter when signing a certificate. Modified so all relevant
3177 operations are affected by the digest parameter including the
3178 -fingerprint and -x509toreq options. Also -x509toreq choked if a
3179 DSA key was used because it didn't fix the digest.
3182 *) Initial certificate chain verify code. Currently tests the untrusted
3183 certificates for consistency with the verify purpose (which is set
3184 when the X509_STORE_CTX structure is set up) and checks the pathlength.
3186 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
3187 this is because it will reject chains with invalid extensions whereas
3188 every previous version of OpenSSL and SSLeay made no checks at all.
3190 Trust code: checks the root CA for the relevant trust settings. Trust
3191 settings have an initial value consistent with the verify purpose: e.g.
3192 if the verify purpose is for SSL client use it expects the CA to be
3193 trusted for SSL client use. However the default value can be changed to
3194 permit custom trust settings: one example of this would be to only trust
3195 certificates from a specific "secure" set of CAs.
3197 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
3198 which should be used for version portability: especially since the
3199 verify structure is likely to change more often now.
3201 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
3202 to set them. If not set then assume SSL clients will verify SSL servers
3205 Two new options to the verify program: -untrusted allows a set of
3206 untrusted certificates to be passed in and -purpose which sets the
3207 intended purpose of the certificate. If a purpose is set then the
3208 new chain verify code is used to check extension consistency.
3211 *) Support for the authority information access extension.
3214 *) Modify RSA and DSA PEM read routines to transparently handle
3215 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
3216 public keys in a format compatible with certificate
3217 SubjectPublicKeyInfo structures. Unfortunately there were already
3218 functions called *_PublicKey_* which used various odd formats so
3219 these are retained for compatibility: however the DSA variants were
3220 never in a public release so they have been deleted. Changed dsa/rsa
3221 utilities to handle the new format: note no releases ever handled public
3222 keys so we should be OK.
3224 The primary motivation for this change is to avoid the same fiasco
3225 that dogs private keys: there are several incompatible private key
3226 formats some of which are standard and some OpenSSL specific and
3227 require various evil hacks to allow partial transparent handling and
3228 even then it doesn't work with DER formats. Given the option anything
3229 other than PKCS#8 should be dumped: but the other formats have to
3230 stay in the name of compatibility.
3232 With public keys and the benefit of hindsight one standard format
3233 is used which works with EVP_PKEY, RSA or DSA structures: though
3234 it clearly returns an error if you try to read the wrong kind of key.
3236 Added a -pubkey option to the 'x509' utility to output the public key.
3237 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
3238 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
3239 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
3240 that do the same as the EVP_PKEY_assign_*() except they up the
3241 reference count of the added key (they don't "swallow" the
3245 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
3246 CRLs would fail if the file contained no certificates or no CRLs:
3247 added a new function to read in both types and return the number
3248 read: this means that if none are read it will be an error. The
3249 DER versions of the certificate and CRL reader would always fail
3250 because it isn't possible to mix certificates and CRLs in DER format
3251 without choking one or the other routine. Changed this to just read
3252 a certificate: this is the best we can do. Also modified the code
3253 in apps/verify.c to take notice of return codes: it was previously
3254 attempting to read in certificates from NULL pointers and ignoring
3255 any errors: this is one reason why the cert and CRL reader seemed
3256 to work. It doesn't check return codes from the default certificate
3257 routines: these may well fail if the certificates aren't installed.
3260 *) Code to support otherName option in GeneralName.
3263 *) First update to verify code. Change the verify utility
3264 so it warns if it is passed a self signed certificate:
3265 for consistency with the normal behaviour. X509_verify
3266 has been modified to it will now verify a self signed
3267 certificate if *exactly* the same certificate appears
3268 in the store: it was previously impossible to trust a
3269 single self signed certificate. This means that:
3270 openssl verify ss.pem
3271 now gives a warning about a self signed certificate but
3272 openssl verify -CAfile ss.pem ss.pem
3276 *) For servers, store verify_result in SSL_SESSION data structure
3277 (and add it to external session representation).
3278 This is needed when client certificate verifications fails,
3279 but an application-provided verification callback (set by
3280 SSL_CTX_set_cert_verify_callback) allows accepting the session
3281 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
3282 but returns 1): When the session is reused, we have to set
3283 ssl->verify_result to the appropriate error code to avoid
3285 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
3287 *) Fix a bug in the new PKCS#7 code: it didn't consider the
3288 case in PKCS7_dataInit() where the signed PKCS7 structure
3289 didn't contain any existing data because it was being created.
3290 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
3292 *) Add a salt to the key derivation routines in enc.c. This
3293 forms the first 8 bytes of the encrypted file. Also add a
3294 -S option to allow a salt to be input on the command line.
3297 *) New function X509_cmp(). Oddly enough there wasn't a function
3298 to compare two certificates. We do this by working out the SHA1
3299 hash and comparing that. X509_cmp() will be needed by the trust
3303 *) SSL_get1_session() is like SSL_get_session(), but increments
3304 the reference count in the SSL_SESSION returned.
3305 [Geoff Thorpe <geoff@eu.c2.net>]
3307 *) Fix for 'req': it was adding a null to request attributes.
3308 Also change the X509_LOOKUP and X509_INFO code to handle
3309 certificate auxiliary information.
3312 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
3316 *) Add the possibility to add extra information to the memory leak
3317 detecting output, to form tracebacks, showing from where each
3318 allocation was originated: CRYPTO_push_info("constant string") adds
3319 the string plus current file name and line number to a per-thread
3320 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
3321 is like calling CYRPTO_pop_info() until the stack is empty.
3322 Also updated memory leak detection code to be multi-thread-safe.
3325 *) Add options -text and -noout to pkcs7 utility and delete the
3326 encryption options which never did anything. Update docs.
3329 *) Add options to some of the utilities to allow the pass phrase
3330 to be included on either the command line (not recommended on
3331 OSes like Unix) or read from the environment. Update the
3332 manpages and fix a few bugs.
3335 *) Add a few manpages for some of the openssl commands.
3338 *) Fix the -revoke option in ca. It was freeing up memory twice,
3339 leaking and not finding already revoked certificates.
3342 *) Extensive changes to support certificate auxiliary information.
3343 This involves the use of X509_CERT_AUX structure and X509_AUX
3344 functions. An X509_AUX function such as PEM_read_X509_AUX()
3345 can still read in a certificate file in the usual way but it
3346 will also read in any additional "auxiliary information". By
3347 doing things this way a fair degree of compatibility can be
3348 retained: existing certificates can have this information added
3349 using the new 'x509' options.
3351 Current auxiliary information includes an "alias" and some trust
3352 settings. The trust settings will ultimately be used in enhanced
3353 certificate chain verification routines: currently a certificate
3354 can only be trusted if it is self signed and then it is trusted
3358 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
3359 The problem was that one of the replacement routines had not been working
3360 since SSLeay releases. For now the offending routine has been replaced
3361 with non-optimised assembler. Even so, this now gives around 95%
3362 performance improvement for 1024 bit RSA signs.
3365 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
3366 handling. Most clients have the effective key size in bits equal to
3367 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
3368 A few however don't do this and instead use the size of the decrypted key
3369 to determine the RC2 key length and the AlgorithmIdentifier to determine
3370 the effective key length. In this case the effective key length can still
3371 be 40 bits but the key length can be 168 bits for example. This is fixed
3372 by manually forcing an RC2 key into the EVP_PKEY structure because the
3373 EVP code can't currently handle unusual RC2 key sizes: it always assumes
3374 the key length and effective key length are equal.
3377 *) Add a bunch of functions that should simplify the creation of
3378 X509_NAME structures. Now you should be able to do:
3379 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
3380 and have it automatically work out the correct field type and fill in
3381 the structures. The more adventurous can try:
3382 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
3383 and it will (hopefully) work out the correct multibyte encoding.
3386 *) Change the 'req' utility to use the new field handling and multibyte
3387 copy routines. Before the DN field creation was handled in an ad hoc
3388 way in req, ca, and x509 which was rather broken and didn't support
3389 BMPStrings or UTF8Strings. Since some software doesn't implement
3390 BMPStrings or UTF8Strings yet, they can be enabled using the config file
3391 using the dirstring_type option. See the new comment in the default
3392 openssl.cnf for more info.
3395 *) Make crypto/rand/md_rand.c more robust:
3396 - Assure unique random numbers after fork().
3397 - Make sure that concurrent threads access the global counter and
3398 md serializably so that we never lose entropy in them
3399 or use exactly the same state in multiple threads.
3400 Access to the large state is not always serializable because
3401 the additional locking could be a performance killer, and
3402 md should be large enough anyway.
3405 *) New file apps/app_rand.c with commonly needed functionality
3406 for handling the random seed file.
3408 Use the random seed file in some applications that previously did not:
3410 dsaparam -genkey (which also ignored its '-rand' option),
3413 x509 (when signing).
3414 Except on systems with /dev/urandom, it is crucial to have a random
3415 seed file at least for key creation, DSA signing, and for DH exchanges;
3416 for RSA signatures we could do without one.
3418 gendh and gendsa (unlike genrsa) used to read only the first byte
3419 of each file listed in the '-rand' option. The function as previously
3420 found in genrsa is now in app_rand.c and is used by all programs
3421 that support '-rand'.
3424 *) In RAND_write_file, use mode 0600 for creating files;
3425 don't just chmod when it may be too late.
3428 *) Report an error from X509_STORE_load_locations
3429 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
3432 *) New function ASN1_mbstring_copy() this copies a string in either
3433 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
3434 into an ASN1_STRING type. A mask of permissible types is passed
3435 and it chooses the "minimal" type to use or an error if not type
3439 *) Add function equivalents to the various macros in asn1.h. The old
3440 macros are retained with an M_ prefix. Code inside the library can
3441 use the M_ macros. External code (including the openssl utility)
3442 should *NOT* in order to be "shared library friendly".
3445 *) Add various functions that can check a certificate's extensions
3446 to see if it usable for various purposes such as SSL client,
3447 server or S/MIME and CAs of these types. This is currently
3448 VERY EXPERIMENTAL but will ultimately be used for certificate chain
3449 verification. Also added a -purpose flag to x509 utility to
3450 print out all the purposes.
3453 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
3457 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
3458 for, obtain and decode and extension and obtain its critical flag.
3459 This allows all the necessary extension code to be handled in a
3460 single function call.
3463 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
3464 platforms. See crypto/rc4/rc4_enc.c for further details.
3467 *) New -noout option to asn1parse. This causes no output to be produced
3468 its main use is when combined with -strparse and -out to extract data
3469 from a file (which may not be in ASN.1 format).
3472 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
3473 when producing the local key id.
3474 [Richard Levitte <levitte@stacken.kth.se>]
3476 *) New option -dhparam in s_server. This allows a DH parameter file to be
3477 stated explicitly. If it is not stated then it tries the first server
3478 certificate file. The previous behaviour hard coded the filename
3482 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
3483 a public key to be input or output. For example:
3484 openssl rsa -in key.pem -pubout -out pubkey.pem
3485 Also added necessary DSA public key functions to handle this.
3488 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
3489 in the message. This was handled by allowing
3490 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
3491 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
3493 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
3494 to the end of the strings whereas this didn't. This would cause problems
3495 if strings read with d2i_ASN1_bytes() were later modified.
3496 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
3498 *) Fix for base64 decode bug. When a base64 bio reads only one line of
3499 data and it contains EOF it will end up returning an error. This is
3500 caused by input 46 bytes long. The cause is due to the way base64
3501 BIOs find the start of base64 encoded data. They do this by trying a
3502 trial decode on each line until they find one that works. When they
3503 do a flag is set and it starts again knowing it can pass all the
3504 data directly through the decoder. Unfortunately it doesn't reset
3505 the context it uses. This means that if EOF is reached an attempt
3506 is made to pass two EOFs through the context and this causes the
3507 resulting error. This can also cause other problems as well. As is
3508 usual with these problems it takes *ages* to find and the fix is
3509 trivial: move one line.
3510 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
3512 *) Ugly workaround to get s_client and s_server working under Windows. The
3513 old code wouldn't work because it needed to select() on sockets and the
3514 tty (for keypresses and to see if data could be written). Win32 only
3515 supports select() on sockets so we select() with a 1s timeout on the
3516 sockets and then see if any characters are waiting to be read, if none
3517 are present then we retry, we also assume we can always write data to
3518 the tty. This isn't nice because the code then blocks until we've
3519 received a complete line of data and it is effectively polling the
3520 keyboard at 1s intervals: however it's quite a bit better than not
3521 working at all :-) A dedicated Windows application might handle this
3522 with an event loop for example.
3525 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
3526 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
3527 will be called when RSA_sign() and RSA_verify() are used. This is useful
3528 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
3529 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
3530 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
3531 This necessitated the support of an extra signature type NID_md5_sha1
3532 for SSL signatures and modifications to the SSL library to use it instead
3533 of calling RSA_public_decrypt() and RSA_private_encrypt().
3536 *) Add new -verify -CAfile and -CApath options to the crl program, these
3537 will lookup a CRL issuers certificate and verify the signature in a
3538 similar way to the verify program. Tidy up the crl program so it
3539 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
3540 less strict. It will now permit CRL extensions even if it is not
3541 a V2 CRL: this will allow it to tolerate some broken CRLs.
3544 *) Initialize all non-automatic variables each time one of the openssl
3545 sub-programs is started (this is necessary as they may be started
3546 multiple times from the "OpenSSL>" prompt).
3547 [Lennart Bang, Bodo Moeller]
3549 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
3550 removing all other RSA functionality (this is what NO_RSA does). This
3551 is so (for example) those in the US can disable those operations covered
3552 by the RSA patent while allowing storage and parsing of RSA keys and RSA
3556 *) Non-copying interface to BIO pairs.
3557 (still largely untested)
3560 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
3561 ASCII string. This was handled independently in various places before.
3564 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
3565 UTF8 strings a character at a time.
3568 *) Use client_version from client hello to select the protocol
3569 (s23_srvr.c) and for RSA client key exchange verification
3570 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
3573 *) Add various utility functions to handle SPKACs, these were previously
3574 handled by poking round in the structure internals. Added new function
3575 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
3576 print, verify and generate SPKACs. Based on an original idea from
3577 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
3580 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
3583 *) Allow the config file extension section to be overwritten on the
3584 command line. Based on an original idea from Massimiliano Pala
3585 <madwolf@comune.modena.it>. The new option is called -extensions
3586 and can be applied to ca, req and x509. Also -reqexts to override
3587 the request extensions in req and -crlexts to override the crl extensions
3591 *) Add new feature to the SPKAC handling in ca. Now you can include
3592 the same field multiple times by preceding it by "XXXX." for example:
3595 this is the same syntax as used in the req config file.
3598 *) Allow certificate extensions to be added to certificate requests. These
3599 are specified in a 'req_extensions' option of the req section of the
3600 config file. They can be printed out with the -text option to req but
3601 are otherwise ignored at present.
3604 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
3605 data read consists of only the final block it would not decrypted because
3606 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
3607 A misplaced 'break' also meant the decrypted final block might not be
3608 copied until the next read.
3611 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
3612 a few extra parameters to the DH structure: these will be useful if
3613 for example we want the value of 'q' or implement X9.42 DH.
3616 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
3617 provides hooks that allow the default DSA functions or functions on a
3618 "per key" basis to be replaced. This allows hardware acceleration and
3619 hardware key storage to be handled without major modification to the
3620 library. Also added low level modexp hooks and CRYPTO_EX structure and
3621 associated functions.
3624 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
3625 as "read only": it can't be written to and the buffer it points to will
3626 not be freed. Reading from a read only BIO is much more efficient than
3627 a normal memory BIO. This was added because there are several times when
3628 an area of memory needs to be read from a BIO. The previous method was
3629 to create a memory BIO and write the data to it, this results in two
3630 copies of the data and an O(n^2) reading algorithm. There is a new
3631 function BIO_new_mem_buf() which creates a read only memory BIO from
3632 an area of memory. Also modified the PKCS#7 routines to use read only
3636 *) Bugfix: ssl23_get_client_hello did not work properly when called in
3637 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
3638 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
3639 but a retry condition occured while trying to read the rest.
3642 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
3643 NID_pkcs7_encrypted by default: this was wrong since this should almost
3644 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
3645 the encrypted data type: this is a more sensible place to put it and it
3646 allows the PKCS#12 code to be tidied up that duplicated this
3650 *) Changed obj_dat.pl script so it takes its input and output files on
3651 the command line. This should avoid shell escape redirection problems
3655 *) Initial support for certificate extension requests, these are included
3656 in things like Xenroll certificate requests. Included functions to allow
3657 extensions to be obtained and added.
3660 *) -crlf option to s_client and s_server for sending newlines as
3661 CRLF (as required by many protocols).
3664 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
3666 *) Install libRSAglue.a when OpenSSL is built with RSAref.
3667 [Ralf S. Engelschall]
3669 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
3670 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
3672 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
3676 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
3677 DH parameters/keys (q is lost during that conversion, but the resulting
3678 DH parameters contain its length).
3680 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
3681 much faster than DH_generate_parameters (which creates parameters
3682 where p = 2*q + 1), and also the smaller q makes DH computations
3683 much more efficient (160-bit exponentiation instead of 1024-bit
3684 exponentiation); so this provides a convenient way to support DHE
3685 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
3686 utter importance to use
3687 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
3689 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
3690 when such DH parameters are used, because otherwise small subgroup
3691 attacks may become possible!
3694 *) Avoid memory leak in i2d_DHparams.
3697 *) Allow the -k option to be used more than once in the enc program:
3698 this allows the same encrypted message to be read by multiple recipients.
3701 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
3702 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
3703 it will always use the numerical form of the OID, even if it has a short
3707 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
3708 method only got called if p,q,dmp1,dmq1,iqmp components were present,
3709 otherwise bn_mod_exp was called. In the case of hardware keys for example
3710 no private key components need be present and it might store extra data
3711 in the RSA structure, which cannot be accessed from bn_mod_exp.
3712 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
3713 private key operations.
3716 *) Added support for SPARC Linux.
3719 *) pem_password_cb function type incompatibly changed from
3720 typedef int pem_password_cb(char *buf, int size, int rwflag);
3722 ....(char *buf, int size, int rwflag, void *userdata);
3723 so that applications can pass data to their callbacks:
3724 The PEM[_ASN1]_{read,write}... functions and macros now take an
3725 additional void * argument, which is just handed through whenever
3726 the password callback is called.
3727 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
3729 New function SSL_CTX_set_default_passwd_cb_userdata.
3731 Compatibility note: As many C implementations push function arguments
3732 onto the stack in reverse order, the new library version is likely to
3733 interoperate with programs that have been compiled with the old
3734 pem_password_cb definition (PEM_whatever takes some data that
3735 happens to be on the stack as its last argument, and the callback
3736 just ignores this garbage); but there is no guarantee whatsoever that
3739 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
3740 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
3741 problems not only on Windows, but also on some Unix platforms.
3742 To avoid problematic command lines, these definitions are now in an
3743 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
3744 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
3747 *) MIPS III/IV assembler module is reimplemented.
3750 *) More DES library cleanups: remove references to srand/rand and
3751 delete an unused file.
3754 *) Add support for the the free Netwide assembler (NASM) under Win32,
3755 since not many people have MASM (ml) and it can be hard to obtain.
3756 This is currently experimental but it seems to work OK and pass all
3757 the tests. Check out INSTALL.W32 for info.
3760 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
3761 without temporary keys kept an extra copy of the server key,
3762 and connections with temporary keys did not free everything in case
3766 *) New function RSA_check_key and new openssl rsa option -check
3767 for verifying the consistency of RSA keys.
3768 [Ulf Moeller, Bodo Moeller]
3770 *) Various changes to make Win32 compile work:
3771 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
3772 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
3773 comparison" warnings.
3774 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
3777 *) Add a debugging option to PKCS#5 v2 key generation function: when
3778 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
3779 derived keys are printed to stderr.
3782 *) Copy the flags in ASN1_STRING_dup().
3783 [Roman E. Pavlov <pre@mo.msk.ru>]
3785 *) The x509 application mishandled signing requests containing DSA
3786 keys when the signing key was also DSA and the parameters didn't match.
3788 It was supposed to omit the parameters when they matched the signing key:
3789 the verifying software was then supposed to automatically use the CA's
3790 parameters if they were absent from the end user certificate.
3792 Omitting parameters is no longer recommended. The test was also
3793 the wrong way round! This was probably due to unusual behaviour in
3794 EVP_cmp_parameters() which returns 1 if the parameters match.
3795 This meant that parameters were omitted when they *didn't* match and
3796 the certificate was useless. Certificates signed with 'ca' didn't have
3798 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
3800 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
3801 The interface is as follows:
3802 Applications can use
3803 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
3804 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
3805 "off" is now the default.
3806 The library internally uses
3807 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
3808 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
3809 to disable memory-checking temporarily.
3811 Some inconsistent states that previously were possible (and were
3812 even the default) are now avoided.
3814 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
3815 with each memory chunk allocated; this is occasionally more helpful
3816 than just having a counter.
3818 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
3820 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
3824 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
3825 which largely parallels "options", but is for changing API behaviour,
3826 whereas "options" are about protocol behaviour.
3827 Initial "mode" flags are:
3829 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
3830 a single record has been written.
3831 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
3832 retries use the same buffer location.
3833 (But all of the contents must be
3837 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
3840 *) Fix problems with no-hmac etc.
3841 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
3843 *) New functions RSA_get_default_method(), RSA_set_method() and
3844 RSA_get_method(). These allows replacement of RSA_METHODs without having
3845 to mess around with the internals of an RSA structure.
3848 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
3849 Also really enable memory leak checks in openssl.c and in some
3851 [Chad C. Mulligan, Bodo Moeller]
3853 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
3854 up the length of negative integers. This has now been simplified to just
3855 store the length when it is first determined and use it later, rather
3856 than trying to keep track of where data is copied and updating it to
3858 [Steve Henson, reported by Brien Wheeler
3859 <bwheeler@authentica-security.com>]
3861 *) Add a new function PKCS7_signatureVerify. This allows the verification
3862 of a PKCS#7 signature but with the signing certificate passed to the
3863 function itself. This contrasts with PKCS7_dataVerify which assumes the
3864 certificate is present in the PKCS#7 structure. This isn't always the
3865 case: certificates can be omitted from a PKCS#7 structure and be
3866 distributed by "out of band" means (such as a certificate database).
3869 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
3870 function prototypes in pem.h, also change util/mkdef.pl to add the
3871 necessary function names.
3874 *) mk1mf.pl (used by Windows builds) did not properly read the
3875 options set by Configure in the top level Makefile, and Configure
3876 was not even able to write more than one option correctly.
3877 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
3880 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
3881 file to be loaded from a BIO or FILE pointer. The BIO version will
3882 for example allow memory BIOs to contain config info.
3885 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
3886 Whoever hopes to achieve shared-library compatibility across versions
3887 must use this, not the compile-time macro.
3888 (Exercise 0.9.4: Which is the minimum library version required by
3890 Note: All this applies only to multi-threaded programs, others don't
3894 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
3895 through a BIO pair triggered the default case, i.e.
3896 SSLerr(...,SSL_R_UNKNOWN_STATE).
3899 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
3900 can use the SSL library even if none of the specific BIOs is
3904 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
3905 for the encoded length.
3906 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
3908 *) Add initial documentation of the X509V3 functions.
3911 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
3912 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
3913 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
3914 secure PKCS#8 private key format with a high iteration count.
3917 *) Fix determination of Perl interpreter: A perl or perl5
3918 _directory_ in $PATH was also accepted as the interpreter.
3919 [Ralf S. Engelschall]
3921 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
3922 wrong with it but it was very old and did things like calling
3923 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
3927 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
3928 to use the new extension code.
3931 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
3932 with macros. This should make it easier to change their form, add extra
3933 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
3937 *) Add to configuration table a new entry that can specify an alternative
3938 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
3939 according to Mark Crispin <MRC@Panda.COM>.
3943 *) DES CBC did not update the IV. Weird.
3946 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
3947 Changing the behaviour of the former might break existing programs --
3948 where IV updating is needed, des_ncbc_encrypt can be used.
3951 *) When bntest is run from "make test" it drives bc to check its
3952 calculations, as well as internally checking them. If an internal check
3953 fails, it needs to cause bc to give a non-zero result or make test carries
3954 on without noticing the failure. Fixed.
3957 *) DES library cleanups.
3960 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
3961 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
3962 ciphers. NOTE: although the key derivation function has been verified
3963 against some published test vectors it has not been extensively tested
3964 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
3968 *) Instead of "mkdir -p", which is not fully portable, use new
3969 Perl script "util/mkdir-p.pl".
3972 *) Rewrite the way password based encryption (PBE) is handled. It used to
3973 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
3974 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
3975 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
3976 the 'parameter' field of the AlgorithmIdentifier is passed to the
3977 underlying key generation function so it must do its own ASN1 parsing.
3978 This has also changed the EVP_PBE_CipherInit() function which now has a
3979 'parameter' argument instead of literal salt and iteration count values
3980 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
3983 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
3984 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
3985 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
3986 KEY" because this clashed with PKCS#8 unencrypted string. Since this
3987 value was just used as a "magic string" and not used directly its
3988 value doesn't matter.
3991 *) Introduce some semblance of const correctness to BN. Shame C doesn't
3995 *) "linux-sparc64" configuration (ultrapenguin).
3996 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
3997 "linux-sparc" configuration.
3998 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
4000 *) config now generates no-xxx options for missing ciphers.
4003 *) Support the EBCDIC character set (work in progress).
4004 File ebcdic.c not yet included because it has a different license.
4005 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
4007 *) Support BS2000/OSD-POSIX.
4008 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
4010 *) Make callbacks for key generation use void * instead of char *.
4013 *) Make S/MIME samples compile (not yet tested).
4016 *) Additional typesafe stacks.
4019 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
4023 Changes between 0.9.3 and 0.9.3a [29 May 1999]
4025 *) New configuration variant "sco5-gcc".
4027 *) Updated some demos.
4028 [Sean O Riordain, Wade Scholine]
4030 *) Add missing BIO_free at exit of pkcs12 application.
4033 *) Fix memory leak in conf.c.
4036 *) Updates for Win32 to assembler version of MD5.
4039 *) Set #! path to perl in apps/der_chop to where we found it
4040 instead of using a fixed path.
4043 *) SHA library changes for irix64-mips4-cc.
4046 *) Improvements for VMS support.
4050 Changes between 0.9.2b and 0.9.3 [24 May 1999]
4052 *) Bignum library bug fix. IRIX 6 passes "make test" now!
4053 This also avoids the problems with SC4.2 and unpatched SC5.
4054 [Andy Polyakov <appro@fy.chalmers.se>]
4056 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
4057 These are required because of the typesafe stack would otherwise break
4058 existing code. If old code used a structure member which used to be STACK
4059 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
4060 sk_num or sk_value it would produce an error because the num, data members
4061 are not present in STACK_OF. Now it just produces a warning. sk_set
4062 replaces the old method of assigning a value to sk_value
4063 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
4064 that does this will no longer work (and should use sk_set instead) but
4065 this could be regarded as a "questionable" behaviour anyway.
4068 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
4069 correctly handle encrypted S/MIME data.
4072 *) Change type of various DES function arguments from des_cblock
4073 (which means, in function argument declarations, pointer to char)
4074 to des_cblock * (meaning pointer to array with 8 char elements),
4075 which allows the compiler to do more typechecking; it was like
4076 that back in SSLeay, but with lots of ugly casts.
4078 Introduce new type const_des_cblock.
4081 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
4082 problems: find RecipientInfo structure that matches recipient certificate
4083 and initialise the ASN1 structures properly based on passed cipher.
4086 *) Belatedly make the BN tests actually check the results.
4089 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
4090 to and from BNs: it was completely broken. New compilation option
4091 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
4092 key elements as negative integers.
4095 *) Reorganize and speed up MD5.
4096 [Andy Polyakov <appro@fy.chalmers.se>]
4099 [Richard Levitte <richard@levitte.org>]
4101 *) New option -out to asn1parse to allow the parsed structure to be
4102 output to a file. This is most useful when combined with the -strparse
4103 option to examine the output of things like OCTET STRINGS.
4106 *) Make SSL library a little more fool-proof by not requiring any longer
4107 that SSL_set_{accept,connect}_state be called before
4108 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
4109 in many applications because usually everything *appeared* to work as
4110 intended anyway -- now it really works as intended).
4113 *) Move openssl.cnf out of lib/.
4116 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
4117 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
4118 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
4119 [Ralf S. Engelschall]
4121 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
4122 handle PKCS#7 enveloped data properly.
4123 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
4125 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
4126 copying pointers. The cert_st handling is changed by this in
4127 various ways (and thus what used to be known as ctx->default_cert
4128 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
4129 any longer when s->cert does not give us what we need).
4130 ssl_cert_instantiate becomes obsolete by this change.
4131 As soon as we've got the new code right (possibly it already is?),
4132 we have solved a couple of bugs of the earlier code where s->cert
4133 was used as if it could not have been shared with other SSL structures.
4135 Note that using the SSL API in certain dirty ways now will result
4136 in different behaviour than observed with earlier library versions:
4137 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
4138 does not influence s as it used to.
4140 In order to clean up things more thoroughly, inside SSL_SESSION
4141 we don't use CERT any longer, but a new structure SESS_CERT
4142 that holds per-session data (if available); currently, this is
4143 the peer's certificate chain and, for clients, the server's certificate
4144 and temporary key. CERT holds only those values that can have
4145 meaningful defaults in an SSL_CTX.
4148 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
4149 from the internal representation. Various PKCS#7 fixes: remove some
4150 evil casts and set the enc_dig_alg field properly based on the signing
4154 *) Allow PKCS#12 password to be set from the command line or the
4155 environment. Let 'ca' get its config file name from the environment
4156 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
4160 *) Allow certificate policies extension to use an IA5STRING for the
4161 organization field. This is contrary to the PKIX definition but
4162 VeriSign uses it and IE5 only recognises this form. Document 'x509'
4166 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
4167 without disallowing inline assembler and the like for non-pedantic builds.
4170 *) Support Borland C++ builder.
4171 [Janez Jere <jj@void.si>, modified by Ulf Möller]
4176 *) SHA-1 cleanups and performance enhancements.
4177 [Andy Polyakov <appro@fy.chalmers.se>]
4179 *) Sparc v8plus assembler for the bignum library.
4180 [Andy Polyakov <appro@fy.chalmers.se>]
4182 *) Accept any -xxx and +xxx compiler options in Configure.
4185 *) Update HPUX configuration.
4188 *) Add missing sk_<type>_unshift() function to safestack.h
4189 [Ralf S. Engelschall]
4191 *) New function SSL_CTX_use_certificate_chain_file that sets the
4192 "extra_cert"s in addition to the certificate. (This makes sense
4193 only for "PEM" format files, as chains as a whole are not
4197 *) Support verify_depth from the SSL API.
4198 x509_vfy.c had what can be considered an off-by-one-error:
4199 Its depth (which was not part of the external interface)
4200 was actually counting the number of certificates in a chain;
4201 now it really counts the depth.
4204 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
4205 instead of X509err, which often resulted in confusing error
4206 messages since the error codes are not globally unique
4207 (e.g. an alleged error in ssl3_accept when a certificate
4208 didn't match the private key).
4210 *) New function SSL_CTX_set_session_id_context that allows to set a default
4211 value (so that you don't need SSL_set_session_id_context for each
4212 connection using the SSL_CTX).
4215 *) OAEP decoding bug fix.
4218 *) Support INSTALL_PREFIX for package builders, as proposed by
4222 *) New Configure options "threads" and "no-threads". For systems
4223 where the proper compiler options are known (currently Solaris
4224 and Linux), "threads" is the default.
4227 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
4230 *) Install various scripts to $(OPENSSLDIR)/misc, not to
4231 $(INSTALLTOP)/bin -- they shouldn't clutter directories
4232 such as /usr/local/bin.
4235 *) "make linux-shared" to build shared libraries.
4236 [Niels Poppe <niels@netbox.org>]
4238 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
4241 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
4242 extension adding in x509 utility.
4245 *) Remove NOPROTO sections and error code comments.
4248 *) Partial rewrite of the DEF file generator to now parse the ANSI
4252 *) New Configure options --prefix=DIR and --openssldir=DIR.
4255 *) Complete rewrite of the error code script(s). It is all now handled
4256 by one script at the top level which handles error code gathering,
4257 header rewriting and C source file generation. It should be much better
4258 than the old method: it now uses a modified version of Ulf's parser to
4259 read the ANSI prototypes in all header files (thus the old K&R definitions
4260 aren't needed for error creation any more) and do a better job of
4261 translating function codes into names. The old 'ASN1 error code imbedded
4262 in a comment' is no longer necessary and it doesn't use .err files which
4263 have now been deleted. Also the error code call doesn't have to appear all
4264 on one line (which resulted in some large lines...).
4267 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
4270 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
4271 0 (which usually indicates a closed connection), but continue reading.
4274 *) Fix some race conditions.
4277 *) Add support for CRL distribution points extension. Add Certificate
4278 Policies and CRL distribution points documentation.
4281 *) Move the autogenerated header file parts to crypto/opensslconf.h.
4284 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
4285 8 of keying material. Merlin has also confirmed interop with this fix
4286 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
4287 [Merlin Hughes <merlin@baltimore.ie>]
4289 *) Fix lots of warnings.
4290 [Richard Levitte <levitte@stacken.kth.se>]
4292 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
4293 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
4294 [Richard Levitte <levitte@stacken.kth.se>]
4296 *) Fix problems with sizeof(long) == 8.
4297 [Andy Polyakov <appro@fy.chalmers.se>]
4299 *) Change functions to ANSI C.
4302 *) Fix typos in error codes.
4303 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
4305 *) Remove defunct assembler files from Configure.
4308 *) SPARC v8 assembler BIGNUM implementation.
4309 [Andy Polyakov <appro@fy.chalmers.se>]
4311 *) Support for Certificate Policies extension: both print and set.
4312 Various additions to support the r2i method this uses.
4315 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
4316 return a const string when you are expecting an allocated buffer.
4319 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
4320 types DirectoryString and DisplayText.
4323 *) Add code to allow r2i extensions to access the configuration database,
4324 add an LHASH database driver and add several ctx helper functions.
4327 *) Fix an evil bug in bn_expand2() which caused various BN functions to
4328 fail when they extended the size of a BIGNUM.
4331 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
4332 support typesafe stack.
4335 *) Fix typo in SSL_[gs]et_options().
4336 [Nils Frostberg <nils@medcom.se>]
4338 *) Delete various functions and files that belonged to the (now obsolete)
4339 old X509V3 handling code.
4342 *) New Configure option "rsaref".
4345 *) Don't auto-generate pem.h.
4348 *) Introduce type-safe ASN.1 SETs.
4351 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
4352 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
4354 *) Introduce type-safe STACKs. This will almost certainly break lots of code
4355 that links with OpenSSL (well at least cause lots of warnings), but fear
4356 not: the conversion is trivial, and it eliminates loads of evil casts. A
4357 few STACKed things have been converted already. Feel free to convert more.
4358 In the fullness of time, I'll do away with the STACK type altogether.
4361 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
4362 specified in <certfile> by updating the entry in the index.txt file.
4363 This way one no longer has to edit the index.txt file manually for
4364 revoking a certificate. The -revoke option does the gory details now.
4365 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
4367 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
4368 `-text' option at all and this way the `-noout -text' combination was
4369 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
4370 [Ralf S. Engelschall]
4372 *) Make sure a corresponding plain text error message exists for the
4373 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
4374 verify callback function determined that a certificate was revoked.
4375 [Ralf S. Engelschall]
4377 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
4378 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
4379 all available cipers including rc5, which was forgotten until now.
4380 In order to let the testing shell script know which algorithms
4381 are available, a new (up to now undocumented) command
4382 "openssl list-cipher-commands" is used.
4385 *) Bugfix: s_client occasionally would sleep in select() when
4386 it should have checked SSL_pending() first.
4389 *) New functions DSA_do_sign and DSA_do_verify to provide access to
4390 the raw DSA values prior to ASN.1 encoding.
4393 *) Tweaks to Configure
4394 [Niels Poppe <niels@netbox.org>]
4396 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
4400 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
4403 *) New config option to avoid instructions that are illegal on the 80386.
4404 The default code is faster, but requires at least a 486.
4407 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
4408 SSL2_SERVER_VERSION (not used at all) macros, which are now the
4409 same as SSL2_VERSION anyway.
4412 *) New "-showcerts" option for s_client.
4415 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
4416 application. Various cleanups and fixes.
4419 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
4420 modify error routines to work internally. Add error codes and PBE init
4421 to library startup routines.
4424 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
4425 packing functions to asn1 and evp. Changed function names and error
4426 codes along the way.
4429 *) PKCS12 integration: and so it begins... First of several patches to
4430 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
4431 objects to objects.h
4434 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
4435 and display support for Thawte strong extranet extension.
4438 *) Add LinuxPPC support.
4439 [Jeff Dubrule <igor@pobox.org>]
4441 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
4442 bn_div_words in alpha.s.
4443 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
4445 *) Make sure the RSA OAEP test is skipped under -DRSAref because
4446 OAEP isn't supported when OpenSSL is built with RSAref.
4447 [Ulf Moeller <ulf@fitug.de>]
4449 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
4450 so they no longer are missing under -DNOPROTO.
4451 [Soren S. Jorvang <soren@t.dk>]
4454 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
4456 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
4457 doesn't work when the session is reused. Coming soon!
4460 *) Fix a security hole, that allows sessions to be reused in the wrong
4461 context thus bypassing client cert protection! All software that uses
4462 client certs and session caches in multiple contexts NEEDS PATCHING to
4463 allow session reuse! A fuller solution is in the works.
4464 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
4466 *) Some more source tree cleanups (removed obsolete files
4467 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
4468 permission on "config" script to be executable) and a fix for the INSTALL
4470 [Ulf Moeller <ulf@fitug.de>]
4472 *) Remove some legacy and erroneous uses of malloc, free instead of
4474 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
4476 *) Make rsa_oaep_test return non-zero on error.
4477 [Ulf Moeller <ulf@fitug.de>]
4479 *) Add support for native Solaris shared libraries. Configure
4480 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
4481 if someone would make that last step automatic.
4482 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
4484 *) ctx_size was not built with the right compiler during "make links". Fixed.
4487 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
4488 except NULL ciphers". This means the default cipher list will no longer
4489 enable NULL ciphers. They need to be specifically enabled e.g. with
4490 the string "DEFAULT:eNULL".
4493 *) Fix to RSA private encryption routines: if p < q then it would
4494 occasionally produce an invalid result. This will only happen with
4495 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
4498 *) Be less restrictive and allow also `perl util/perlpath.pl
4499 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
4500 because this way one can also use an interpreter named `perl5' (which is
4501 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
4502 installed as `perl').
4503 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
4505 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
4506 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
4508 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
4509 advapi32.lib to Win32 build and change the pem test comparision
4510 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
4511 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
4512 and crypto/des/ede_cbcm_enc.c.
4515 *) DES quad checksum was broken on big-endian architectures. Fixed.
4518 *) Comment out two functions in bio.h that aren't implemented. Fix up the
4519 Win32 test batch file so it (might) work again. The Win32 test batch file
4520 is horrible: I feel ill....
4523 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
4524 in e_os.h. Audit of header files to check ANSI and non ANSI
4525 sections: 10 functions were absent from non ANSI section and not exported
4526 from Windows DLLs. Fixed up libeay.num for new functions.
4529 *) Make `openssl version' output lines consistent.
4530 [Ralf S. Engelschall]
4532 *) Fix Win32 symbol export lists for BIO functions: Added
4533 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
4534 to ms/libeay{16,32}.def.
4535 [Ralf S. Engelschall]
4537 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
4538 fine under Unix and passes some trivial tests I've now added. But the
4539 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
4540 added to make sure no one expects that this stuff really works in the
4541 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
4542 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
4544 [Ralf S. Engelschall]
4546 *) Fix the generation of two part addresses in perl.
4547 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
4549 *) Add config entry for Linux on MIPS.
4550 [John Tobey <jtobey@channel1.com>]
4552 *) Make links whenever Configure is run, unless we are on Windoze.
4555 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
4556 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
4560 *) Add a useful kludge to allow package maintainers to specify compiler and
4561 other platforms details on the command line without having to patch the
4562 Configure script everytime: One now can use ``perl Configure
4563 <id>:<details>'', i.e. platform ids are allowed to have details appended
4564 to them (seperated by colons). This is treated as there would be a static
4565 pre-configured entry in Configure's %table under key <id> with value
4566 <details> and ``perl Configure <id>'' is called. So, when you want to
4567 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
4568 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
4569 now, which overrides the FreeBSD-elf entry on-the-fly.
4570 [Ralf S. Engelschall]
4572 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
4575 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
4576 on the `perl Configure ...' command line. This way one can compile
4577 OpenSSL libraries with Position Independent Code (PIC) which is needed
4578 for linking it into DSOs.
4579 [Ralf S. Engelschall]
4581 *) Remarkably, export ciphers were totally broken and no-one had noticed!
4585 *) Cleaned up the LICENSE document: The official contact for any license
4586 questions now is the OpenSSL core team under openssl-core@openssl.org.
4587 And add a paragraph about the dual-license situation to make sure people
4588 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
4589 to the OpenSSL toolkit.
4590 [Ralf S. Engelschall]
4592 *) General source tree makefile cleanups: Made `making xxx in yyy...'
4593 display consistent in the source tree and replaced `/bin/rm' by `rm'.
4594 Additonally cleaned up the `make links' target: Remove unnecessary
4595 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
4596 to speed processing and no longer clutter the display with confusing
4597 stuff. Instead only the actually done links are displayed.
4598 [Ralf S. Engelschall]
4600 *) Permit null encryption ciphersuites, used for authentication only. It used
4601 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
4602 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
4606 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
4607 signed attributes when verifying signatures (this would break them),
4608 the detached data encoding was wrong and public keys obtained using
4609 X509_get_pubkey() weren't freed.
4612 *) Add text documentation for the BUFFER functions. Also added a work around
4613 to a Win95 console bug. This was triggered by the password read stuff: the
4614 last character typed gets carried over to the next fread(). If you were
4615 generating a new cert request using 'req' for example then the last
4616 character of the passphrase would be CR which would then enter the first
4620 *) Added the new `Includes OpenSSL Cryptography Software' button as
4621 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
4622 button and can be used by applications based on OpenSSL to show the
4623 relationship to the OpenSSL project.
4624 [Ralf S. Engelschall]
4626 *) Remove confusing variables in function signatures in files
4627 ssl/ssl_lib.c and ssl/ssl.h.
4628 [Lennart Bong <lob@kulthea.stacken.kth.se>]
4630 *) Don't install bss_file.c under PREFIX/include/
4631 [Lennart Bong <lob@kulthea.stacken.kth.se>]
4633 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
4634 functions that return function pointers and has support for NT specific
4635 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
4636 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
4637 unsigned to signed types: this was killing the Win32 compile.
4640 *) Add new certificate file to stack functions,
4641 SSL_add_dir_cert_subjects_to_stack() and
4642 SSL_add_file_cert_subjects_to_stack(). These largely supplant
4643 SSL_load_client_CA_file(), and can be used to add multiple certs easily
4644 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
4645 This means that Apache-SSL and similar packages don't have to mess around
4646 to add as many CAs as they want to the preferred list.
4649 *) Experiment with doxygen documentation. Currently only partially applied to
4651 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
4652 openssl.doxy as the configuration file.
4655 *) Get rid of remaining C++-style comments which strict C compilers hate.
4656 [Ralf S. Engelschall, pointed out by Carlos Amengual]
4658 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
4659 compiled in by default: it has problems with large keys.
4662 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
4663 DH private keys and/or callback functions which directly correspond to
4664 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
4665 is needed for applications which have to configure certificates on a
4666 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
4668 For the RSA certificate situation is makes no difference, but
4669 for the DSA certificate situation this fixes the "no shared cipher"
4670 problem where the OpenSSL cipher selection procedure failed because the
4671 temporary keys were not overtaken from the context and the API provided
4672 no way to reconfigure them.
4673 The new functions now let applications reconfigure the stuff and they
4674 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
4675 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
4676 non-public-API function ssl_cert_instantiate() is used as a helper
4677 function and also to reduce code redundancy inside ssl_rsa.c.
4678 [Ralf S. Engelschall]
4680 *) Move s_server -dcert and -dkey options out of the undocumented feature
4681 area because they are useful for the DSA situation and should be
4682 recognized by the users.
4683 [Ralf S. Engelschall]
4685 *) Fix the cipher decision scheme for export ciphers: the export bits are
4686 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
4687 SSL_EXP_MASK. So, the original variable has to be used instead of the
4688 already masked variable.
4689 [Richard Levitte <levitte@stacken.kth.se>]
4691 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
4692 [Richard Levitte <levitte@stacken.kth.se>]
4694 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
4695 from `int' to `unsigned int' because it's a length and initialized by
4696 EVP_DigestFinal() which expects an `unsigned int *'.
4697 [Richard Levitte <levitte@stacken.kth.se>]
4699 *) Don't hard-code path to Perl interpreter on shebang line of Configure
4700 script. Instead use the usual Shell->Perl transition trick.
4701 [Ralf S. Engelschall]
4703 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
4704 (in addition to RSA certificates) to match the behaviour of `openssl dsa
4705 -noout -modulus' as it's already the case for `openssl rsa -noout
4706 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
4707 currently the public key is printed (a decision which was already done by
4708 `openssl dsa -modulus' in the past) which serves a similar purpose.
4709 Additionally the NO_RSA no longer completely removes the whole -modulus
4710 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
4712 [Ralf S. Engelschall]
4714 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
4715 BIO. See the source (crypto/evp/bio_ok.c) for more info.
4716 [Arne Ansper <arne@ats.cyber.ee>]
4718 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
4719 to be added. Now both 'req' and 'ca' can use new objects defined in the
4723 *) Add cool BIO that does syslog (or event log on NT).
4724 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
4726 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
4727 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
4728 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
4729 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
4732 *) Add preliminary config info for new extension code.
4735 *) Make RSA_NO_PADDING really use no padding.
4736 [Ulf Moeller <ulf@fitug.de>]
4738 *) Generate errors when private/public key check is done.
4741 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
4742 for some CRL extensions and new objects added.
4745 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
4746 key usage extension and fuller support for authority key id.
4749 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
4750 padding method for RSA, which is recommended for new applications in PKCS
4751 #1 v2.0 (RFC 2437, October 1998).
4752 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
4753 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
4754 against Bleichbacher's attack on RSA.
4755 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
4758 *) Updates to the new SSL compression code
4759 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
4761 *) Fix so that the version number in the master secret, when passed
4762 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
4763 (because the server will not accept higher), that the version number
4764 is 0x03,0x01, not 0x03,0x00
4765 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
4767 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
4768 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
4769 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
4772 *) Support for RAW extensions where an arbitrary extension can be
4773 created by including its DER encoding. See apps/openssl.cnf for
4777 *) Make sure latest Perl versions don't interpret some generated C array
4778 code as Perl array code in the crypto/err/err_genc.pl script.
4779 [Lars Weber <3weber@informatik.uni-hamburg.de>]
4781 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
4782 not many people have the assembler. Various Win32 compilation fixes and
4783 update to the INSTALL.W32 file with (hopefully) more accurate Win32
4787 *) Modify configure script 'Configure' to automatically create crypto/date.h
4788 file under Win32 and also build pem.h from pem.org. New script
4789 util/mkfiles.pl to create the MINFO file on environments that can't do a
4790 'make files': perl util/mkfiles.pl >MINFO should work.
4793 *) Major rework of DES function declarations, in the pursuit of correctness
4794 and purity. As a result, many evil casts evaporated, and some weirdness,
4795 too. You may find this causes warnings in your code. Zapping your evil
4796 casts will probably fix them. Mostly.
4799 *) Fix for a typo in asn1.h. Bug fix to object creation script
4800 obj_dat.pl. It considered a zero in an object definition to mean
4801 "end of object": none of the objects in objects.h have any zeros
4802 so it wasn't spotted.
4803 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
4805 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
4806 Masking (CBCM). In the absence of test vectors, the best I have been able
4807 to do is check that the decrypt undoes the encrypt, so far. Send me test
4808 vectors if you have them.
4811 *) Correct calculation of key length for export ciphers (too much space was
4812 allocated for null ciphers). This has not been tested!
4815 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
4816 message is now correct (it understands "crypto" and "ssl" on its
4817 command line). There is also now an "update" option. This will update
4818 the util/ssleay.num and util/libeay.num files with any new functions.
4820 perl util/mkdef.pl crypto ssl update
4821 it will update them.
4824 *) Overhauled the Perl interface (perl/*):
4825 - ported BN stuff to OpenSSL's different BN library
4826 - made the perl/ source tree CVS-aware
4827 - renamed the package from SSLeay to OpenSSL (the files still contain
4828 their history because I've copied them in the repository)
4829 - removed obsolete files (the test scripts will be replaced
4830 by better Test::Harness variants in the future)
4831 [Ralf S. Engelschall]
4833 *) First cut for a very conservative source tree cleanup:
4834 1. merge various obsolete readme texts into doc/ssleay.txt
4835 where we collect the old documents and readme texts.
4836 2. remove the first part of files where I'm already sure that we no
4837 longer need them because of three reasons: either they are just temporary
4838 files which were left by Eric or they are preserved original files where
4839 I've verified that the diff is also available in the CVS via "cvs diff
4840 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
4841 the crypto/md/ stuff).
4842 [Ralf S. Engelschall]
4844 *) More extension code. Incomplete support for subject and issuer alt
4845 name, issuer and authority key id. Change the i2v function parameters
4846 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
4847 what that's for :-) Fix to ASN1 macro which messed up
4848 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
4851 *) Preliminary support for ENUMERATED type. This is largely copied from the
4855 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
4856 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
4858 *) Make sure `make rehash' target really finds the `openssl' program.
4859 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
4861 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
4862 like to hear about it if this slows down other processors.
4865 *) Add CygWin32 platform information to Configure script.
4866 [Alan Batie <batie@aahz.jf.intel.com>]
4868 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
4869 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
4871 *) New program nseq to manipulate netscape certificate sequences
4874 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
4878 *) Fixes to BN code. Previously the default was to define BN_RECURSION
4879 but the BN code had some problems that would cause failures when
4880 doing certificate verification and some other functions.
4881 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
4883 *) Add ASN1 and PEM code to support netscape certificate sequences.
4886 *) Add ASN1 and PEM code to support netscape certificate sequences.
4889 *) Add several PKIX and private extended key usage OIDs.
4892 *) Modify the 'ca' program to handle the new extension code. Modify
4893 openssl.cnf for new extension format, add comments.
4896 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
4897 and add a sample to openssl.cnf so req -x509 now adds appropriate
4901 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
4902 error code, add initial support to X509_print() and x509 application.
4905 *) Takes a deep breath and start addding X509 V3 extension support code. Add
4906 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
4907 stuff is currently isolated and isn't even compiled yet.
4910 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
4911 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
4912 Removed the versions check from X509 routines when loading extensions:
4913 this allows certain broken certificates that don't set the version
4914 properly to be processed.
4917 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
4918 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
4919 can still be regenerated with "make depend".
4922 *) Spelling mistake in C version of CAST-128.
4923 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
4925 *) Changes to the error generation code. The perl script err-code.pl
4926 now reads in the old error codes and retains the old numbers, only
4927 adding new ones if necessary. It also only changes the .err files if new
4928 codes are added. The makefiles have been modified to only insert errors
4929 when needed (to avoid needlessly modifying header files). This is done
4930 by only inserting errors if the .err file is newer than the auto generated
4931 C file. To rebuild all the error codes from scratch (the old behaviour)
4932 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
4933 or delete all the .err files.
4936 *) CAST-128 was incorrectly implemented for short keys. The C version has
4937 been fixed, but is untested. The assembler versions are also fixed, but
4938 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
4939 to regenerate it if needed.
4940 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
4941 Hagino <itojun@kame.net>]
4943 *) File was opened incorrectly in randfile.c.
4944 [Ulf Möller <ulf@fitug.de>]
4946 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
4947 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
4948 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
4949 al: it's just almost always a UTCTime. Note this patch adds new error
4950 codes so do a "make errors" if there are problems.
4953 *) Correct Linux 1 recognition in config.
4954 [Ulf Möller <ulf@fitug.de>]
4956 *) Remove pointless MD5 hash when using DSA keys in ca.
4957 [Anonymous <nobody@replay.com>]
4959 *) Generate an error if given an empty string as a cert directory. Also
4960 generate an error if handed NULL (previously returned 0 to indicate an
4961 error, but didn't set one).
4962 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
4964 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
4967 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
4968 parameters. This was causing a warning which killed off the Win32 compile.
4971 *) Remove C++ style comments from crypto/bn/bn_local.h.
4972 [Neil Costigan <neil.costigan@celocom.com>]
4974 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
4975 based on a text string, looking up short and long names and finally
4976 "dot" format. The "dot" format stuff didn't work. Added new function
4977 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
4978 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
4979 OID is not part of the table.
4982 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
4983 X509_LOOKUP_by_alias().
4986 *) Sort openssl functions by name.
4989 *) Get the gendsa program working (hopefully) and add it to app list. Remove
4990 encryption from sample DSA keys (in case anyone is interested the password
4994 *) Make _all_ *_free functions accept a NULL pointer.
4995 [Frans Heymans <fheymans@isaserver.be>]
4997 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
4999 [Anonymous <nobody@replay.com>]
5001 *) s_server should send the CAfile as acceptable CAs, not its own cert.
5002 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
5004 *) Don't blow it for numeric -newkey arguments to apps/req.
5005 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
5007 *) Temp key "for export" tests were wrong in s3_srvr.c.
5008 [Anonymous <nobody@replay.com>]
5010 *) Add prototype for temp key callback functions
5011 SSL_CTX_set_tmp_{rsa,dh}_callback().
5014 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
5015 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
5018 *) X509_name_add_entry() freed the wrong thing after an error.
5019 [Arne Ansper <arne@ats.cyber.ee>]
5021 *) rsa_eay.c would attempt to free a NULL context.
5022 [Arne Ansper <arne@ats.cyber.ee>]
5024 *) BIO_s_socket() had a broken should_retry() on Windoze.
5025 [Arne Ansper <arne@ats.cyber.ee>]
5027 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
5028 [Arne Ansper <arne@ats.cyber.ee>]
5030 *) Make sure the already existing X509_STORE->depth variable is initialized
5031 in X509_STORE_new(), but document the fact that this variable is still
5032 unused in the certificate verification process.
5033 [Ralf S. Engelschall]
5035 *) Fix the various library and apps files to free up pkeys obtained from
5036 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
5039 *) Fix reference counting in X509_PUBKEY_get(). This makes
5040 demos/maurice/example2.c work, amongst others, probably.
5041 [Steve Henson and Ben Laurie]
5043 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
5044 `openssl' and second, the shortcut symlinks for the `openssl <command>'
5045 are no longer created. This way we have a single and consistent command
5046 line interface `openssl <command>', similar to `cvs <command>'.
5047 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
5049 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
5050 BIT STRING wrapper always have zero unused bits.
5053 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
5056 *) Make the top-level INSTALL documentation easier to understand.
5059 *) Makefiles updated to exit if an error occurs in a sub-directory
5060 make (including if user presses ^C) [Paul Sutton]
5062 *) Make Montgomery context stuff explicit in RSA data structure.
5065 *) Fix build order of pem and err to allow for generated pem.h.
5068 *) Fix renumbering bug in X509_NAME_delete_entry().
5071 *) Enhanced the err-ins.pl script so it makes the error library number
5072 global and can add a library name. This is needed for external ASN1 and
5073 other error libraries.
5076 *) Fixed sk_insert which never worked properly.
5079 *) Fix ASN1 macros so they can handle indefinite length construted
5080 EXPLICIT tags. Some non standard certificates use these: they can now
5084 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
5085 into a single doc/ssleay.txt bundle. This way the information is still
5086 preserved but no longer messes up this directory. Now it's new room for
5087 the new set of documenation files.
5088 [Ralf S. Engelschall]
5090 *) SETs were incorrectly DER encoded. This was a major pain, because they
5091 shared code with SEQUENCEs, which aren't coded the same. This means that
5092 almost everything to do with SETs or SEQUENCEs has either changed name or
5093 number of arguments.
5094 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
5096 *) Fix test data to work with the above.
5099 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
5100 was already fixed by Eric for 0.9.1 it seems.
5101 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
5103 *) Autodetect FreeBSD3.
5106 *) Fix various bugs in Configure. This affects the following platforms:
5110 unixware-2.0-pentium
5114 *) Eliminate generated files from CVS. Reorder tests to regenerate files
5115 before they are needed.
5118 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
5122 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
5124 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
5125 changed SSLeay to OpenSSL in version strings.
5126 [Ralf S. Engelschall]
5128 *) Some fixups to the top-level documents.
5131 *) Fixed the nasty bug where rsaref.h was not found under compile-time
5132 because the symlink to include/ was missing.
5133 [Ralf S. Engelschall]
5135 *) Incorporated the popular no-RSA/DSA-only patches
5136 which allow to compile a RSA-free SSLeay.
5137 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
5139 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
5140 when "ssleay" is still not found.
5141 [Ralf S. Engelschall]
5143 *) Added more platforms to Configure: Cray T3E, HPUX 11,
5144 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
5146 *) Updated the README file.
5147 [Ralf S. Engelschall]
5149 *) Added various .cvsignore files in the CVS repository subdirs
5150 to make a "cvs update" really silent.
5151 [Ralf S. Engelschall]
5153 *) Recompiled the error-definition header files and added
5154 missing symbols to the Win32 linker tables.
5155 [Ralf S. Engelschall]
5157 *) Cleaned up the top-level documents;
5158 o new files: CHANGES and LICENSE
5159 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
5160 o merged COPYRIGHT into LICENSE
5161 o removed obsolete TODO file
5162 o renamed MICROSOFT to INSTALL.W32
5163 [Ralf S. Engelschall]
5165 *) Removed dummy files from the 0.9.1b source tree:
5166 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
5167 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
5168 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
5169 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
5170 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
5171 [Ralf S. Engelschall]
5173 *) Added various platform portability fixes.
5176 *) The Genesis of the OpenSSL rpject:
5177 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
5178 Young and Tim J. Hudson created while they were working for C2Net until
5180 [The OpenSSL Project]
5183 Changes between 0.9.0b and 0.9.1b [not released]
5185 *) Updated a few CA certificates under certs/
5188 *) Changed some BIGNUM api stuff.
5191 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
5192 DGUX x86, Linux Alpha, etc.
5195 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
5196 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
5200 *) Add -strparse option to asn1pars program which parses nested
5202 [Dr Stephen Henson <shenson@bigfoot.com>]
5204 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
5207 *) DSA fix for "ca" program.
5210 *) Added "-genkey" option to "dsaparam" program.
5213 *) Added RIPE MD160 (rmd160) message digest.
5216 *) Added -a (all) option to "ssleay version" command.
5219 *) Added PLATFORM define which is the id given to Configure.
5222 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
5225 *) Extended the ASN.1 parser routines.
5228 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
5231 *) Added a BN_CTX to the BN library.
5234 *) Fixed the weak key values in DES library
5237 *) Changed API in EVP library for cipher aliases.
5240 *) Added support for RC2/64bit cipher.
5243 *) Converted the lhash library to the crypto/mem.c functions.
5246 *) Added more recognized ASN.1 object ids.
5249 *) Added more RSA padding checks for SSL/TLS.
5252 *) Added BIO proxy/filter functionality.
5255 *) Added extra_certs to SSL_CTX which can be used
5256 send extra CA certificates to the client in the CA cert chain sending
5257 process. It can be configured with SSL_CTX_add_extra_chain_cert().
5260 *) Now Fortezza is denied in the authentication phase because
5261 this is key exchange mechanism is not supported by SSLeay at all.
5264 *) Additional PKCS1 checks.
5267 *) Support the string "TLSv1" for all TLS v1 ciphers.
5270 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
5271 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
5274 *) Fixed a few memory leaks.
5277 *) Fixed various code and comment typos.
5280 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
5281 bytes sent in the client random.
5282 [Edward Bishop <ebishop@spyglass.com>]