5 Changes between 0.9.6 and 0.9.7 [xx XXX 2001]
7 OpenSSL 0.9.6a/0.9.6b (bugfix releases, 5 Apr 2001 and 9 July 2001)
8 and OpenSSL 0.9.7 were developed in parallel, based on OpenSSL 0.9.6.
10 Change log entries are tagged as follows:
11 -) applies to 0.9.6a/0.9.6b/0.9.6c only
12 *) applies to 0.9.6a/0.9.6b/0.9.6c and 0.9.7
13 +) applies to 0.9.7 only
15 +) Add support for shared libraries for Unixware-7 and support including
16 shared libraries for OpenUNIX-8 (Boyd Lynn Gerber <gerberb@zenez.com>).
19 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
20 'wristwatch attack' using huge encoding parameters (cf.
21 James H. Manger's CRYPTO 2001 paper). Note that the
22 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
23 encoding parameters and hence was not vulnerable.
26 +) Add a "destroy" handler to ENGINEs that allows structural cleanup to
27 be done prior to destruction. Use this to unload error strings from
28 ENGINEs that load their own error strings. NB: This adds two new API
29 functions to "get" and "set" this destroy handler in an ENGINE.
32 +) Alter all existing ENGINE implementations (except "openssl" and
33 "openbsd") to dynamically instantiate their own error strings. This
34 makes them more flexible to be built both as statically-linked ENGINEs
35 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
36 Also, add stub code to each that makes building them as self-contained
37 shared-libraries easier (see README.ENGINE).
40 +) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
41 implementations into applications that are completely implemented in
42 self-contained shared-libraries. The "dynamic" ENGINE exposes control
43 commands that can be used to configure what shared-library to load and
44 to control aspects of the way it is handled. Also, made an update to
45 the README.ENGINE file that brings its information up-to-date and
46 provides some information and instructions on the "dynamic" ENGINE
47 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
51 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
53 *) Make it possible to unload ranges of ERR strings with a new
54 "ERR_unload_strings" function.
57 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
58 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
59 followed by modular reduction.
60 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
62 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
63 equivalent based on BN_pseudo_rand() instead of BN_rand().
66 +) Add a copy() function to EVP_MD.
69 +) Make EVP_MD routines take a context pointer instead of just the
73 +) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
74 that the digest can only process a single chunk of data
75 (typically because it is provided by a piece of
76 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
77 is only going to provide a single chunk of data, and hence the
78 framework needn't accumulate the data for oneshot drivers.
81 +) As with "ERR", make it possible to replace the underlying "ex_data"
82 functions. This change also alters the storage and management of global
83 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
84 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
85 index counters. The API functions that use this state have been changed
86 to take a "class_index" rather than pointers to the class's local STACK
87 and counter, and there is now an API function to dynamically create new
88 classes. This centralisation allows us to (a) plug a lot of the
89 thread-safety problems that existed, and (b) makes it possible to clean
90 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
91 such data would previously have always leaked in application code and
92 workarounds were in place to make the memory debugging turn a blind eye
93 to it. Application code that doesn't use this new function will still
94 leak as before, but their memory debugging output will announce it now
95 rather than letting it slide.
97 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
98 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
99 has a return value to indicate success or failure.
102 +) Make it possible to replace the underlying "ERR" functions such that the
103 global state (2 LHASH tables and 2 locks) is only used by the "default"
104 implementation. This change also adds two functions to "get" and "set"
105 the implementation prior to it being automatically set the first time
106 any other ERR function takes place. Ie. an application can call "get",
107 pass the return value to a module it has just loaded, and that module
108 can call its own "set" function using that value. This means the
109 module's "ERR" operations will use (and modify) the error state in the
110 application and not in its own statically linked copy of OpenSSL code.
113 +) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
114 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
115 the operation, and provides a more encapsulated way for external code
116 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
117 to use these functions rather than manually incrementing the counts.
119 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
122 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
123 This function was broken, as the check for a new client hello message
124 to handle SGC did not allow these large messages.
125 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
128 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
131 +) Add EVP test program.
134 +) Add symmetric cipher support to ENGINE. Expect the API to change!
137 +) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
138 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
139 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
140 These allow a CRL to be built without having to access X509_CRL fields
141 directly. Modify 'ca' application to use new functions.
144 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
145 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
148 *) Rework the configuration and shared library support for Tru64 Unix.
149 The configuration part makes use of modern compiler features and
150 still retains old compiler behavior for those that run older versions
151 of the OS. The shared library support part includes a variant that
152 uses the RPATH feature, and is available through the special
153 configuration target "alpha-cc-rpath", which will never be selected
155 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
157 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
158 with the same message size as in ssl3_get_certificate_request().
159 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
160 messages might inadvertently be reject as too long.
161 [Petr Lampa <lampa@fee.vutbr.cz>]
163 +) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
164 bug workarounds. Rollback attack detection is a security feature.
165 The problem will only arise on OpenSSL servers when TLSv1 is not
166 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
167 Software authors not wanting to support TLSv1 will have special reasons
168 for their choice and can explicitly enable this option.
169 [Bodo Moeller, Lutz Jaenicke]
171 +) Rationalise EVP so it can be extended: don't include a union of
172 cipher/digest structures, add init/cleanup functions. This also reduces
173 the number of header dependencies.
176 +) Make DES key schedule conform to the usual scheme, as well as
177 correcting its structure. This means that calls to DES functions
178 now have to pass a pointer to a des_key_schedule instead of a
179 plain des_key_schedule (which was actually always a pointer
183 +) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
186 *) Modified SSL library such that the verify_callback that has been set
187 specificly for an SSL object with SSL_set_verify() is actually being
188 used. Before the change, a verify_callback set with this function was
189 ignored and the verify_callback() set in the SSL_CTX at the time of
190 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
191 to allow the necessary settings.
194 +) Initial reduction of linker bloat: the use of some functions, such as
195 PEM causes large amounts of unused functions to be linked in due to
196 poor organisation. For example pem_all.c contains every PEM function
197 which has a knock on effect of linking in large amounts of (unused)
198 ASN1 code. Grouping together similar functions and splitting unrelated
199 functions prevents this.
202 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
203 explicitly to NULL, as at least on Solaris 8 this seems not always to be
204 done automatically (in contradiction to the requirements of the C
205 standard). This made problems when used from OpenSSH.
208 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
209 dh->length and always used
211 BN_rand_range(priv_key, dh->p).
213 BN_rand_range() is not necessary for Diffie-Hellman, and this
214 specific range makes Diffie-Hellman unnecessarily inefficient if
215 dh->length (recommended exponent length) is much smaller than the
216 length of dh->p. We could use BN_rand_range() if the order of
217 the subgroup was stored in the DH structure, but we only have
222 BN_rand(priv_key, l, ...)
224 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
230 RSA_eay_public_encrypt
231 RSA_eay_private_decrypt
232 RSA_eay_private_encrypt (signing)
233 RSA_eay_public_decrypt (signature verification)
235 (default implementations for RSA_public_encrypt,
236 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
237 always reject numbers >= n.
240 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
241 to synchronize access to 'locking_thread'. This is necessary on
242 systems where access to 'locking_thread' (an 'unsigned long'
243 variable) is not atomic.
246 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
247 *before* setting the 'crypto_lock_rand' flag. The previous code had
248 a race condition if 0 is a valid thread ID.
249 [Travis Vitek <vitek@roguewave.com>]
251 +) Cleanup of EVP macros.
254 +) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
258 +) Add initial OCSP responder support to ocsp application. The
259 revocation information is handled using the text based index
260 use by the ca application. The responder can either handle
261 requests generated internally, supplied in files (for example
262 via a CGI script) or using an internal minimal server.
265 +) Add configuration choices to get zlib compression for TLS.
268 +) Changes to Kerberos SSL for RFC 2712 compliance:
269 1. Implemented real KerberosWrapper, instead of just using
270 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
271 2. Implemented optional authenticator field of KerberosWrapper.
273 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
274 and authenticator structs; see crypto/krb5/.
276 Generalized Kerberos calls to support multiple Kerberos libraries.
277 [Vern Staats <staatsvr@asc.hpc.mil>,
278 Jeffrey Altman <jaltman@columbia.edu>
281 +) Cause 'openssl speed' to use fully hard-coded DSA keys as it
282 already does with RSA. testdsa.h now has 'priv_key/pub_key'
283 values for each of the key sizes rather than having just
284 parameters (and 'speed' generating keys each time).
287 -) OpenSSL 0.9.6b released [9 July 2001]
289 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
290 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
291 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
292 PRNG state recovery was possible based on the output of
293 one PRNG request appropriately sized to gain knowledge on
294 'md' followed by enough consecutive 1-byte PRNG requests
295 to traverse all of 'state'.
297 1. When updating 'md_local' (the current thread's copy of 'md')
298 during PRNG output generation, hash all of the previous
299 'md_local' value, not just the half used for PRNG output.
301 2. Make the number of bytes from 'state' included into the hash
302 independent from the number of PRNG bytes requested.
304 The first measure alone would be sufficient to avoid
305 Markku-Juhani's attack. (Actually it had never occurred
306 to me that the half of 'md_local' used for chaining was the
307 half from which PRNG output bytes were taken -- I had always
308 assumed that the secret half would be used.) The second
309 measure makes sure that additional data from 'state' is never
310 mixed into 'md_local' in small portions; this heuristically
311 further strengthens the PRNG.
314 +) Speed up EVP routines.
317 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
318 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
319 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
320 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
322 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
323 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
324 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
327 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
329 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
332 *) Fix crypto/bn/asm/mips3.s.
335 *) When only the key is given to "enc", the IV is undefined. Print out
336 an error message in this case.
339 +) Added the OS2-EMX target.
340 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
342 +) Rewrite apps to use NCONF routines instead of the old CONF. New functions
343 to support NCONF routines in extension code. New function CONF_set_nconf()
344 to allow functions which take an NCONF to also handle the old LHASH
345 structure: this means that the old CONF compatible routines can be
346 retained (in particular wrt extensions) without having to duplicate the
347 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
350 *) Handle special case when X509_NAME is empty in X509 printing routines.
353 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
354 positive and less than q.
357 +) Enhance the general user interface with mechanisms for inner control
358 and with possibilities to have yes/no kind of prompts.
361 +) Change all calls to low level digest routines in the library and
362 applications to use EVP. Add missing calls to HMAC_cleanup() and
363 don't assume HMAC_CTX can be copied using memcpy().
364 [Verdon Walker <VWalker@novell.com>, Steve Henson]
366 +) Add the possibility to control engines through control names but with
367 arbitrary arguments instead of just a string.
368 Change the key loaders to take a UI_METHOD instead of a callback
369 function pointer. NOTE: this breaks binary compatibility with earlier
370 versions of OpenSSL [engine].
371 Adapt the nCipher code for these new conditions and add a card insertion
375 +) Enhance the general user interface with mechanisms to better support
376 dialog box interfaces, application-defined prompts, the possibility
377 to use defaults (for example default passwords from somewhere else)
378 and interrupts/cancellations.
381 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
382 used: it isn't thread safe and the add_lock_callback should handle
384 [Paul Rose <Paul.Rose@bridge.com>]
386 *) Verify that incoming data obeys the block size in
387 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
390 +) Tidy up PKCS#12 attribute handling. Add support for the CSP name
391 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
395 [Ulf Möller, Bodo Möller]
397 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
398 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
399 when fixing the server behaviour for backwards-compatible 'client
400 hello' messages. (Note that the attack is impractical against
401 SSL 3.0 and TLS 1.0 anyway because length and version checking
402 means that the probability of guessing a valid ciphertext is
403 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
406 Before 0.9.5, the countermeasure (hide the error by generating a
407 random 'decryption result') did not work properly because
408 ERR_clear_error() was missing, meaning that SSL_get_error() would
409 detect the supposedly ignored error.
411 Both problems are now fixed.
414 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
415 (previously it was 1024).
418 +) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
419 tidy up some unnecessarily weird code in 'sk_new()').
420 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
422 +) Change the key loading routines for ENGINEs to use the same kind
423 callback (pem_password_cb) as all other routines that need this
427 *) Fix for compatibility mode trust settings: ignore trust settings
428 unless some valid trust or reject settings are present.
431 *) Fix for blowfish EVP: its a variable length cipher.
434 +) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
435 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
436 than this minimum value is recommended.
439 +) New random seeder for OpenVMS, using the system process statistics
440 that are easily reachable.
443 +) Windows apparently can't transparently handle global
444 variables defined in DLLs. Initialisations such as:
446 const ASN1_ITEM *it = &ASN1_INTEGER_it;
448 wont compile. This is used by the any applications that need to
449 declare their own ASN1 modules. This was fixed by adding the option
450 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
451 needed for static libraries under Win32.
454 +) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
455 setting of purpose and trust fields. New X509_STORE trust and
456 purpose functions and tidy up setting in other SSL functions.
459 +) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
460 structure. These are inherited by X509_STORE_CTX when it is
461 initialised. This allows various defaults to be set in the
462 X509_STORE structure (such as flags for CRL checking and custom
463 purpose or trust settings) for functions which only use X509_STORE_CTX
464 internally such as S/MIME.
466 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
467 trust settings if they are not set in X509_STORE. This allows X509_STORE
468 purposes and trust (in S/MIME for example) to override any set by default.
470 Add command line options for CRL checking to smime, s_client and s_server
474 +) Initial CRL based revocation checking. If the CRL checking flag(s)
475 are set then the CRL is looked up in the X509_STORE structure and
476 its validity and signature checked, then if the certificate is found
477 in the CRL the verify fails with a revoked error.
479 Various new CRL related callbacks added to X509_STORE_CTX structure.
481 Command line options added to 'verify' application to support this.
483 This needs some additional work, such as being able to handle multiple
484 CRLs with different times, extension based lookup (rather than just
485 by subject name) and ultimately more complete V2 CRL extension
489 +) Add a general user interface API (crypto/ui/). This is designed
490 to replace things like des_read_password and friends (backward
491 compatibility functions using this new API are provided).
492 The purpose is to remove prompting functions from the DES code
493 section as well as provide for prompting through dialog boxes in
494 a window system and the like.
497 *) Fix various bugs related to DSA S/MIME verification. Handle missing
498 parameters in DSA public key structures and return an error in the
499 DSA routines if parameters are absent.
502 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
503 in the current directory if neither $RANDFILE nor $HOME was set.
504 RAND_file_name() in 0.9.6a returned NULL in this case. This has
505 caused some confusion to Windows users who haven't defined $HOME.
506 Thus RAND_file_name() is changed again: e_os.h can define a
507 DEFAULT_HOME, which will be used if $HOME is not set.
508 For Windows, we use "C:"; on other platforms, we still require
509 environment variables.
511 +) Add "ex_data" support to ENGINE so implementations can add state at a
512 per-structure level rather than having to store it globally.
515 +) Make it possible for ENGINE structures to be copied when retrieved by
516 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
517 This causes the "original" ENGINE structure to act like a template,
518 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
519 operational state can be localised to each ENGINE structure, despite the
520 fact they all share the same "methods". New ENGINE structures returned in
521 this case have no functional references and the return value is the single
522 structural reference. This matches the single structural reference returned
523 by ENGINE_by_id() normally, when it is incremented on the pre-existing
527 +) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
528 needs to match any other type at all we need to manually clear the
532 +) Changes to the "openssl engine" utility to include;
533 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
534 about an ENGINE's available control commands.
535 - executing control commands from command line arguments using the
536 '-pre' and '-post' switches. '-post' is only used if '-t' is
537 specified and the ENGINE is successfully initialised. The syntax for
538 the individual commands are colon-separated, for example;
539 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
542 +) New dynamic control command support for ENGINEs. ENGINEs can now
543 declare their own commands (numbers), names (strings), descriptions,
544 and input types for run-time discovery by calling applications. A
545 subset of these commands are implicitly classed as "executable"
546 depending on their input type, and only these can be invoked through
547 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
548 can be based on user input, config files, etc). The distinction is
549 that "executable" commands cannot return anything other than a boolean
550 result and can only support numeric or string input, whereas some
551 discoverable commands may only be for direct use through
552 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
553 pointers, or other custom uses. The "executable" commands are to
554 support parameterisations of ENGINE behaviour that can be
555 unambiguously defined by ENGINEs and used consistently across any
556 OpenSSL-based application. Commands have been added to all the
557 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
558 control over shared-library paths without source code alterations.
561 +) Changed all ENGINE implementations to dynamically allocate their
562 ENGINEs rather than declaring them statically. Apart from this being
563 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
564 this also allows the implementations to compile without using the
565 internal engine_int.h header.
568 +) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
569 'const' value. Any code that should be able to modify a RAND_METHOD
570 should already have non-const pointers to it (ie. they should only
571 modify their own ones).
574 +) Made a variety of little tweaks to the ENGINE code.
575 - "atalla" and "ubsec" string definitions were moved from header files
576 to C code. "nuron" string definitions were placed in variables
577 rather than hard-coded - allowing parameterisation of these values
578 later on via ctrl() commands.
579 - Removed unused "#if 0"'d code.
580 - Fixed engine list iteration code so it uses ENGINE_free() to release
581 structural references.
582 - Constified the RAND_METHOD element of ENGINE structures.
583 - Constified various get/set functions as appropriate and added
584 missing functions (including a catch-all ENGINE_cpy that duplicates
585 all ENGINE values onto a new ENGINE except reference counts/state).
586 - Removed NULL parameter checks in get/set functions. Setting a method
587 or function to NULL is a way of canceling out a previously set
588 value. Passing a NULL ENGINE parameter is just plain stupid anyway
589 and doesn't justify the extra error symbols and code.
590 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
591 flags from engine_int.h to engine.h.
592 - Changed prototypes for ENGINE handler functions (init(), finish(),
593 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
596 *) Move 'if (!initialized) RAND_poll()' into regions protected by
597 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
598 having multiple threads call RAND_poll() concurrently.
601 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
602 combination of a flag and a thread ID variable.
603 Otherwise while one thread is in ssleay_rand_bytes (which sets the
604 flag), *other* threads can enter ssleay_add_bytes without obeying
605 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
606 that they do not hold after the first thread unsets add_do_not_lock).
609 +) Implement binary inversion algorithm for BN_mod_inverse in addition
610 to the algorithm using long division. The binary algorithm can be
611 used only if the modulus is odd. On 32-bit systems, it is faster
612 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
613 roughly 5-15% for 256-bit moduli), so we use it only for moduli
614 up to 450 bits. In 64-bit environments, the binary algorithm
615 appears to be advantageous for much longer moduli; here we use it
616 for moduli up to 2048 bits.
619 *) Change bctest again: '-x' expressions are not available in all
623 -) OpenSSL 0.9.6a released [5 Apr 2001]
625 *) Fix a couple of memory leaks in PKCS7_dataDecode()
626 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
628 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
629 the default extension for executables, if any. Also, make the perl
630 scripts that use symlink() to test if it really exists and use "cp"
631 if it doesn't. All this made OpenSSL compilable and installable in
635 +) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
636 could not support the combine flag in choice fields.
639 -) Fix for asn1_GetSequence() for indefinite length constructed data.
640 If SEQUENCE is length is indefinite just set c->slen to the total
641 amount of data available.
642 [Steve Henson, reported by shige@FreeBSD.org]
643 [This change does not apply to 0.9.7.]
645 *) Change bctest to avoid here-documents inside command substitution
646 (workaround for FreeBSD /bin/sh bug).
647 For compatibility with Ultrix, avoid shell functions (introduced
648 in the bctest version that searches along $PATH).
651 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
652 with des_encrypt() defined on some operating systems, like Solaris
656 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
657 On the Importance of Eliminating Errors in Cryptographic
658 Computations, J. Cryptology 14 (2001) 2, 101-119,
659 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
662 *) MIPS assembler BIGNUM division bug fix.
665 *) Disabled incorrect Alpha assembler code.
668 -) Fix PKCS#7 decode routines so they correctly update the length
669 after reading an EOC for the EXPLICIT tag.
671 [This change does not apply to 0.9.7.]
673 *) Fix bug in PKCS#12 key generation routines. This was triggered
674 if a 3DES key was generated with a 0 initial byte. Include
675 PKCS12_BROKEN_KEYGEN compilation option to retain the old
676 (but broken) behaviour.
679 *) Enhance bctest to search for a working bc along $PATH and print
681 [Tim Rice <tim@multitalents.net> via Richard Levitte]
683 +) Add a 'copy_extensions' option to the 'ca' utility. This copies
684 extensions from a certificate request to the certificate.
687 +) Allow multiple 'certopt' and 'nameopt' options to be separated
688 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
689 file: this allows the display of the certificate about to be
690 signed to be customised, to allow certain fields to be included
691 or excluded and extension details. The old system didn't display
692 multicharacter strings properly, omitted fields not in the policy
693 and couldn't display additional details such as extensions.
696 *) Fix memory leaks in err.c: free err_data string if necessary;
697 don't write to the wrong index in ERR_set_error_data.
700 +) Function EC_POINTs_mul for simultaneous scalar multiplication
701 of an arbitrary number of elliptic curve points, optionally
702 including the generator defined for the EC_GROUP.
703 EC_POINT_mul is a simple wrapper function for the typical case
704 that the point list has just one item (besides the optional
708 +) First EC_METHODs for curves over GF(p):
710 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
711 operations and provides various method functions that can also
712 operate with faster implementations of modular arithmetic.
714 EC_GFp_mont_method() reuses most functions that are part of
715 EC_GFp_simple_method, but uses Montgomery arithmetic.
717 [Bodo Moeller; point addition and point doubling
718 implementation directly derived from source code provided by
719 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
721 +) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
724 Curves are EC_GROUP objects (with an optional group generator)
725 based on EC_METHODs that are built into the library.
727 Points are EC_POINT objects based on EC_GROUP objects.
729 Most of the framework would be able to handle curves over arbitrary
730 finite fields, but as there are no obvious types for fields other
731 than GF(p), some functions are limited to that for now.
734 +) Add the -HTTP option to s_server. It is similar to -WWW, but requires
735 that the file contains a complete HTTP response.
738 +) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
739 change the def and num file printf format specifier from "%-40sXXX"
740 to "%-39s XXX". The latter will always guarantee a space after the
741 field while the former will cause them to run together if the field
742 is 40 of more characters long.
745 +) Constify the cipher and digest 'method' functions and structures
746 and modify related functions to take constant EVP_MD and EVP_CIPHER
750 *) Implement ssl23_peek (analogous to ssl23_read), which previously
754 *) Replace rdtsc with _emit statements for VC++ version 5.
755 [Jeremy Cooper <jeremy@baymoo.org>]
757 +) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
758 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
761 +) Modify EVP_Digest*() routines so they now return values. Although the
762 internal software routines can never fail additional hardware versions
766 +) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
768 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
769 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
772 ERR_R_NESTED_ASN1_ERROR
774 ERR_R_MISSING_ASN1_EOS
775 were 4 .. 9, conflicting with
776 ERR_LIB_RSA (= ERR_R_RSA_LIB)
778 ERR_LIB_PEM (= ERR_R_PEM_LIB).
779 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
781 Add new error code 'ERR_R_INTERNAL_ERROR'.
784 +) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
788 +) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
789 sets the subject name for a new request or supersedes the
790 subject name in a given request. Formats that can be parsed are
791 'CN=Some Name, OU=myOU, C=IT'
793 'CN=Some Name/OU=myOU/C=IT'.
795 Add options '-batch' and '-verbose' to 'openssl req'.
796 [Massimiliano Pala <madwolf@hackmasters.net>]
798 *) Make it possible to reuse SSLv2 sessions.
801 +) Introduce the possibility to access global variables through
802 functions on platform were that's the best way to handle exporting
803 global variables in shared libraries. To enable this functionality,
804 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
805 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
806 is normally done by Configure or something similar).
808 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
809 in the source file (foo.c) like this:
811 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
812 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
814 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
815 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
817 OPENSSL_DECLARE_GLOBAL(int,foo);
818 #define foo OPENSSL_GLOBAL_REF(foo)
819 OPENSSL_DECLARE_GLOBAL(double,bar);
820 #define bar OPENSSL_GLOBAL_REF(bar)
822 The #defines are very important, and therefore so is including the
823 header file everywhere where the defined globals are used.
825 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
826 of ASN.1 items, but that structure is a bit different.
828 The largest change is in util/mkdef.pl which has been enhanced with
829 better and easier to understand logic to choose which symbols should
830 go into the Windows .def files as well as a number of fixes and code
831 cleanup (among others, algorithm keywords are now sorted
832 lexicographically to avoid constant rewrites).
835 *) In copy_email() check for >= 0 as a return value for
836 X509_NAME_get_index_by_NID() since 0 is a valid index.
837 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
839 +) In BN_div() keep a copy of the sign of 'num' before writing the
840 result to 'rm' because if rm==num the value will be overwritten
841 and produce the wrong result if 'num' is negative: this caused
842 problems with BN_mod() and BN_nnmod().
845 +) Function OCSP_request_verify(). This checks the signature on an
846 OCSP request and verifies the signer certificate. The signer
847 certificate is just checked for a generic purpose and OCSP request
851 +) Add OCSP_check_validity() function to check the validity of OCSP
852 responses. OCSP responses are prepared in real time and may only
853 be a few seconds old. Simply checking that the current time lies
854 between thisUpdate and nextUpdate max reject otherwise valid responses
855 caused by either OCSP responder or client clock inaccuracy. Instead
856 we allow thisUpdate and nextUpdate to fall within a certain period of
857 the current time. The age of the response can also optionally be
858 checked. Two new options -validity_period and -status_age added to
862 +) If signature or public key algorithm is unrecognized print out its
863 OID rather that just UNKNOWN.
866 *) Avoid coredump with unsupported or invalid public keys by checking if
867 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
868 PKCS7_verify() fails with non detached data.
871 +) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
872 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
873 ID to be generated from the issuer certificate alone which can then be
874 passed to OCSP_id_issuer_cmp().
877 +) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
878 ASN1 modules to export functions returning ASN1_ITEM pointers
879 instead of the ASN1_ITEM structures themselves. This adds several
880 new macros which allow the underlying ASN1 function/structure to
881 be accessed transparently. As a result code should not use ASN1_ITEM
882 references directly (such as &X509_it) but instead use the relevant
883 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
884 use of the new ASN1 code on platforms where exporting structures
885 is problematical (for example in shared libraries) but exporting
886 functions returning pointers to structures is not.
889 +) Add support for overriding the generation of SSL/TLS session IDs.
890 These callbacks can be registered either in an SSL_CTX or per SSL.
891 The purpose of this is to allow applications to control, if they wish,
892 the arbitrary values chosen for use as session IDs, particularly as it
893 can be useful for session caching in multiple-server environments. A
894 command-line switch for testing this (and any client code that wishes
895 to use such a feature) has been added to "s_server".
896 [Geoff Thorpe, Lutz Jaenicke]
898 +) Modify mkdef.pl to recognise and parse preprocessor conditionals
899 of the form '#if defined(...) || defined(...) || ...' and
900 '#if !defined(...) && !defined(...) && ...'. This also avoids
901 the growing number of special cases it was previously handling.
904 *) Don't use getenv in library functions when run as setuid/setgid.
905 New function OPENSSL_issetugid().
908 +) Make all configuration macros available for application by making
909 sure they are available in opensslconf.h, by giving them names starting
910 with "OPENSSL_" to avoid conflicts with other packages and by making
911 sure e_os2.h will cover all platform-specific cases together with
913 Additionally, it is now possible to define configuration/platform-
914 specific names (called "system identities"). In the C code, these
915 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
916 macro with the name beginning with "OPENSSL_SYS_", which is determined
917 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
921 +) New option -set_serial to 'req' and 'x509' this allows the serial
922 number to use to be specified on the command line. Previously self
923 signed certificates were hard coded with serial number 0 and the
924 CA options of 'x509' had to use a serial number in a file which was
928 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
929 due to incorrect handling of multi-threading:
931 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
933 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
935 3. Count how many times MemCheck_off() has been called so that
936 nested use can be treated correctly. This also avoids
937 inband-signalling in the previous code (which relied on the
938 assumption that thread ID 0 is impossible).
941 +) New options to 'ca' utility to support V2 CRL entry extensions.
942 Currently CRL reason, invalidity date and hold instruction are
943 supported. Add new CRL extensions to V3 code and some new objects.
946 *) Add "-rand" option also to s_client and s_server.
949 +) New function EVP_CIPHER_CTX_set_padding() this is used to
950 disable standard block padding (aka PKCS#5 padding) in the EVP
951 API, which was previously mandatory. This means that the data is
952 not padded in any way and so the total length much be a multiple
953 of the block size, otherwise an error occurs.
956 +) Initial (incomplete) OCSP SSL support.
959 *) Fix CPU detection on Irix 6.x.
960 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
961 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
963 +) New function OCSP_parse_url(). This splits up a URL into its host,
964 port and path components: primarily to parse OCSP URLs. New -url
965 option to ocsp utility.
968 +) New nonce behavior. The return value of OCSP_check_nonce() now
969 reflects the various checks performed. Applications can decide
970 whether to tolerate certain situations such as an absent nonce
971 in a response when one was present in a request: the ocsp application
972 just prints out a warning. New function OCSP_add1_basic_nonce()
973 this is to allow responders to include a nonce in a response even if
974 the request is nonce-less.
977 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
980 [This change does not apply to 0.9.7.]
982 *) Use the cached encoding of an X509_NAME structure rather than
983 copying it. This is apparently the reason for the libsafe "errors"
984 but the code is actually correct.
987 +) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
988 skipped when using openssl x509 multiple times on a single input file,
989 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
992 +) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
993 set string type: to handle setting ASN1_TIME structures. Fix ca
994 utility to correctly initialize revocation date of CRLs.
997 +) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
998 the clients preferred ciphersuites and rather use its own preferences.
999 Should help to work around M$ SGC (Server Gated Cryptography) bug in
1000 Internet Explorer by ensuring unchanged hash method during stepup.
1001 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
1004 +) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
1005 to aes and add a new 'exist' option to print out symbols that don't
1009 +) Additional options to ocsp utility to allow flags to be set and
1010 additional certificates supplied.
1013 +) Add the option -VAfile to 'openssl ocsp', so the user can give the
1014 OCSP client a number of certificate to only verify the response
1018 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
1019 Bleichenbacher's DSA attack.
1020 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
1021 to be set and top=0 forces the highest bit to be set; top=-1 is new
1022 and leaves the highest bit random.
1023 [Ulf Moeller, Bodo Moeller]
1025 +) Update Rijndael code to version 3.0 and change EVP AES ciphers to
1026 handle the new API. Currently only ECB, CBC modes supported. Add new
1027 AES OIDs. Add TLS AES ciphersuites as described in the "AES Ciphersuites
1028 for TLS" draft-ietf-tls-ciphersuite-03.txt.
1029 [Ben Laurie, Steve Henson]
1031 *) In the NCONF_...-based implementations for CONF_... queries
1032 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
1033 a temporary CONF structure with the data component set to NULL
1034 (which gives segmentation faults in lh_retrieve).
1035 Instead, use NULL for the CONF pointer in CONF_get_string and
1036 CONF_get_number (which may use environment variables) and directly
1037 return NULL from CONF_get_section.
1040 *) Fix potential buffer overrun for EBCDIC.
1043 +) New function OCSP_copy_nonce() to copy nonce value (if present) from
1044 request to response.
1047 +) Functions for OCSP responders. OCSP_request_onereq_count(),
1048 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
1049 extract information from a certificate request. OCSP_response_create()
1050 creates a response and optionally adds a basic response structure.
1051 OCSP_basic_add1_status() adds a complete single response to a basic
1052 response and returns the OCSP_SINGLERESP structure just added (to allow
1053 extensions to be included for example). OCSP_basic_add1_cert() adds a
1054 certificate to a basic response and OCSP_basic_sign() signs a basic
1055 response with various flags. New helper functions ASN1_TIME_check()
1056 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
1057 (converts ASN1_TIME to GeneralizedTime).
1060 +) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
1061 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
1062 structure from a certificate. X509_pubkey_digest() digests the public_key
1063 contents: this is used in various key identifiers.
1066 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
1067 keyUsage if basicConstraints absent for a CA.
1070 *) Make SMIME_write_PKCS7() write mail header values with a format that
1071 is more generally accepted (no spaces before the semicolon), since
1072 some programs can't parse those values properly otherwise. Also make
1073 sure BIO's that break lines after each write do not create invalid
1077 +) Make sk_sort() tolerate a NULL argument.
1078 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
1080 +) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
1081 passed by the function are trusted implicitly. If any of them signed the
1082 response then it is assumed to be valid and is not verified.
1085 -) Make the CRL encoding routines work with empty SEQUENCE OF. The
1086 macros previously used would not encode an empty SEQUENCE OF
1087 and break the signature.
1089 [This change does not apply to 0.9.7.]
1091 *) Zero the premaster secret after deriving the master secret in
1095 +) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
1096 to data. This was previously part of the PKCS7 ASN1 code. This
1097 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
1098 [Steve Henson, reported by Kenneth R. Robinette
1099 <support@securenetterm.com>]
1101 +) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
1102 routines: without these tracing memory leaks is very painful.
1103 Fix leaks in PKCS12 and PKCS7 routines.
1106 *) Add some EVP_add_digest_alias registrations (as found in
1107 OpenSSL_add_all_digests()) to SSL_library_init()
1108 aka OpenSSL_add_ssl_algorithms(). This provides improved
1109 compatibility with peers using X.509 certificates
1110 with unconventional AlgorithmIdentifier OIDs.
1113 *) Fix for Irix with NO_ASM.
1114 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
1116 *) ./config script fixes.
1117 [Ulf Moeller, Richard Levitte]
1119 +) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
1120 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
1121 effectively meant GeneralizedTime would never be used. Now it
1122 is initialised to -1 but X509_time_adj() now has to check the value
1123 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
1124 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
1125 [Steve Henson, reported by Kenneth R. Robinette
1126 <support@securenetterm.com>]
1128 +) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
1129 result in a zero length in the ASN1_INTEGER structure which was
1130 not consistent with the structure when d2i_ASN1_INTEGER() was used
1131 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
1132 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
1133 where it did not print out a minus for negative ASN1_INTEGER.
1136 *) Fix 'openssl passwd -1'.
1139 +) Add summary printout to ocsp utility. The various functions which
1140 convert status values to strings have been renamed to:
1141 OCSP_response_status_str(), OCSP_cert_status_str() and
1142 OCSP_crl_reason_str() and are no longer static. New options
1143 to verify nonce values and to disable verification. OCSP response
1144 printout format cleaned up.
1147 +) Add additional OCSP certificate checks. These are those specified
1148 in RFC2560. This consists of two separate checks: the CA of the
1149 certificate being checked must either be the OCSP signer certificate
1150 or the issuer of the OCSP signer certificate. In the latter case the
1151 OCSP signer certificate must contain the OCSP signing extended key
1152 usage. This check is performed by attempting to match the OCSP
1153 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
1154 in the OCSP_CERTID structures of the response.
1157 +) Initial OCSP certificate verification added to OCSP_basic_verify()
1158 and related routines. This uses the standard OpenSSL certificate
1159 verify routines to perform initial checks (just CA validity) and
1160 to obtain the certificate chain. Then additional checks will be
1161 performed on the chain. Currently the root CA is checked to see
1162 if it is explicitly trusted for OCSP signing. This is used to set
1163 a root CA as a global signing root: that is any certificate that
1164 chains to that CA is an acceptable OCSP signing certificate.
1167 +) New '-extfile ...' option to 'openssl ca' for reading X.509v3
1168 extensions from a separate configuration file.
1169 As when reading extensions from the main configuration file,
1170 the '-extensions ...' option may be used for specifying the
1172 [Massimiliano Pala <madwolf@comune.modena.it>]
1174 *) Change PKCS12_key_gen_asc() so it can cope with non null
1175 terminated strings whose length is passed in the passlen
1176 parameter, for example from PEM callbacks. This was done
1177 by adding an extra length parameter to asc2uni().
1178 [Steve Henson, reported by <oddissey@samsung.co.kr>]
1180 +) New OCSP utility. Allows OCSP requests to be generated or
1181 read. The request can be sent to a responder and the output
1182 parsed, outputed or printed in text form. Not complete yet:
1183 still needs to check the OCSP response validity.
1186 +) New subcommands for 'openssl ca':
1187 'openssl ca -status <serial>' prints the status of the cert with
1188 the given serial number (according to the index file).
1189 'openssl ca -updatedb' updates the expiry status of certificates
1191 [Massimiliano Pala <madwolf@comune.modena.it>]
1193 +) New '-newreq-nodes' command option to CA.pl. This is like
1194 '-newreq', but calls 'openssl req' with the '-nodes' option
1195 so that the resulting key is not encrypted.
1196 [Damien Miller <djm@mindrot.org>]
1198 +) New configuration for the GNU Hurd.
1199 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
1201 +) Initial code to implement OCSP basic response verify. This
1202 is currently incomplete. Currently just finds the signer's
1203 certificate and verifies the signature on the response.
1206 +) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
1207 value of OPENSSLDIR. This is available via the new '-d' option
1208 to 'openssl version', and is also included in 'openssl version -a'.
1211 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
1212 call failed, free the DSA structure.
1215 +) Allowing defining memory allocation callbacks that will be given
1216 file name and line number information in additional arguments
1217 (a const char* and an int). The basic functionality remains, as
1218 well as the original possibility to just replace malloc(),
1219 realloc() and free() by functions that do not know about these
1220 additional arguments. To register and find out the current
1221 settings for extended allocation functions, the following
1222 functions are provided:
1224 CRYPTO_set_mem_ex_functions
1225 CRYPTO_set_locked_mem_ex_functions
1226 CRYPTO_get_mem_ex_functions
1227 CRYPTO_get_locked_mem_ex_functions
1229 These work the same way as CRYPTO_set_mem_functions and friends.
1230 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
1231 extended allocation function is enabled.
1232 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
1233 a conventional allocation function is enabled.
1234 [Richard Levitte, Bodo Moeller]
1236 *) Fix to uni2asc() to cope with zero length Unicode strings.
1237 These are present in some PKCS#12 files.
1240 +) Finish off removing the remaining LHASH function pointer casts.
1241 There should no longer be any prototype-casting required when using
1242 the LHASH abstraction, and any casts that remain are "bugs". See
1243 the callback types and macros at the head of lhash.h for details
1244 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
1247 +) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
1248 If /dev/[u]random devices are not available or do not return enough
1249 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
1251 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
1252 /etc/entropy will be queried once each in this sequence, quering stops
1253 when enough entropy was collected without querying more sockets.
1256 +) Change the Unix RAND_poll() variant to be able to poll several
1257 random devices, as specified by DEVRANDOM, until a sufficient amount
1258 of data has been collected. We spend at most 10 ms on each file
1259 (select timeout) and read in non-blocking mode. DEVRANDOM now
1260 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
1261 (previously it was just the string "/dev/urandom"), so on typical
1262 platforms the 10 ms delay will never occur.
1263 Also separate out the Unix variant to its own file, rand_unix.c.
1264 For VMS, there's a currently-empty rand_vms.c.
1267 +) Move OCSP client related routines to ocsp_cl.c. These
1268 provide utility functions which an application needing
1269 to issue a request to an OCSP responder and analyse the
1270 response will typically need: as opposed to those which an
1271 OCSP responder itself would need which will be added later.
1273 OCSP_request_sign() signs an OCSP request with an API similar
1274 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
1275 response. OCSP_response_get1_basic() extracts basic response
1276 from response. OCSP_resp_find_status(): finds and extracts status
1277 information from an OCSP_CERTID structure (which will be created
1278 when the request structure is built). These are built from lower
1279 level functions which work on OCSP_SINGLERESP structures but
1280 wont normally be used unless the application wishes to examine
1281 extensions in the OCSP response for example.
1283 Replace nonce routines with a pair of functions.
1284 OCSP_request_add1_nonce() adds a nonce value and optionally
1285 generates a random value. OCSP_check_nonce() checks the
1286 validity of the nonce in an OCSP response.
1289 +) Change function OCSP_request_add() to OCSP_request_add0_id().
1290 This doesn't copy the supplied OCSP_CERTID and avoids the
1291 need to free up the newly created id. Change return type
1292 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
1293 This can then be used to add extensions to the request.
1294 Deleted OCSP_request_new(), since most of its functionality
1295 is now in OCSP_REQUEST_new() (and the case insensitive name
1296 clash) apart from the ability to set the request name which
1297 will be added elsewhere.
1300 +) Update OCSP API. Remove obsolete extensions argument from
1301 various functions. Extensions are now handled using the new
1302 OCSP extension code. New simple OCSP HTTP function which
1303 can be used to send requests and parse the response.
1306 +) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
1307 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
1308 uses the special reorder version of SET OF to sort the attributes
1309 and reorder them to match the encoded order. This resolves a long
1310 standing problem: a verify on a PKCS7 structure just after signing
1311 it used to fail because the attribute order did not match the
1312 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
1313 it uses the received order. This is necessary to tolerate some broken
1314 software that does not order SET OF. This is handled by encoding
1315 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
1316 to produce the required SET OF.
1319 +) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
1320 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
1321 files to get correct declarations of the ASN.1 item variables.
1324 +) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
1325 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
1326 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
1327 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
1328 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
1329 ASN1_ITEM and no wrapper functions.
1332 +) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
1333 replace the old function pointer based I/O routines. Change most of
1334 the *_d2i_bio() and *_d2i_fp() functions to use these.
1337 +) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
1338 lines, recognice more "algorithms" that can be deselected, and make
1339 it complain about algorithm deselection that isn't recognised.
1342 +) New ASN1 functions to handle dup, sign, verify, digest, pack and
1343 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
1344 to use new functions. Add NO_ASN1_OLD which can be set to remove
1345 some old style ASN1 functions: this can be used to determine if old
1346 code will still work when these eventually go away.
1349 +) New extension functions for OCSP structures, these follow the
1350 same conventions as certificates and CRLs.
1353 +) New function X509V3_add1_i2d(). This automatically encodes and
1354 adds an extension. Its behaviour can be customised with various
1355 flags to append, replace or delete. Various wrappers added for
1356 certifcates and CRLs.
1359 +) Fix to avoid calling the underlying ASN1 print routine when
1360 an extension cannot be parsed. Correct a typo in the
1361 OCSP_SERVICELOC extension. Tidy up print OCSP format.
1364 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
1365 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
1366 when writing a 32767 byte record.
1367 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
1369 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
1370 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
1372 (RSA objects have a reference count access to which is protected
1373 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
1374 so they are meant to be shared between threads.)
1375 [Bodo Moeller, Geoff Thorpe; original patch submitted by
1376 "Reddie, Steven" <Steven.Reddie@ca.com>]
1378 +) Make mkdef.pl parse some of the ASN1 macros and add apropriate
1379 entries for variables.
1382 *) Fix a deadlock in CRYPTO_mem_leaks().
1385 +) Add functionality to apps/openssl.c for detecting locking
1386 problems: As the program is single-threaded, all we have
1387 to do is register a locking callback using an array for
1388 storing which locks are currently held by the program.
1391 +) Use a lock around the call to CRYPTO_get_ex_new_index() in
1392 SSL_get_ex_data_X509_STORE_idx(), which is used in
1393 ssl_verify_cert_chain() and thus can be called at any time
1394 during TLS/SSL handshakes so that thread-safety is essential.
1395 Unfortunately, the ex_data design is not at all suited
1396 for multi-threaded use, so it probably should be abolished.
1399 +) Added Broadcom "ubsec" ENGINE to OpenSSL.
1400 [Broadcom, tweaked and integrated by Geoff Thorpe]
1402 +) Move common extension printing code to new function
1403 X509V3_print_extensions(). Reorganise OCSP print routines and
1404 implement some needed OCSP ASN1 functions. Add OCSP extensions.
1407 +) New function X509_signature_print() to remove duplication in some
1411 +) Add a special meaning when SET OF and SEQUENCE OF flags are both
1412 set (this was treated exactly the same as SET OF previously). This
1413 is used to reorder the STACK representing the structure to match the
1414 encoding. This will be used to get round a problem where a PKCS7
1415 structure which was signed could not be verified because the STACK
1416 order did not reflect the encoded order.
1419 +) Reimplement the OCSP ASN1 module using the new code.
1422 +) Update the X509V3 code to permit the use of an ASN1_ITEM structure
1423 for its ASN1 operations. The old style function pointers still exist
1424 for now but they will eventually go away.
1427 +) Merge in replacement ASN1 code from the ASN1 branch. This almost
1428 completely replaces the old ASN1 functionality with a table driven
1429 encoder and decoder which interprets an ASN1_ITEM structure describing
1430 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
1431 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
1432 has also been converted to the new form.
1435 +) Change BN_mod_exp_recp so that negative moduli are tolerated
1436 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
1437 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
1438 for negative moduli.
1441 +) Fix BN_uadd and BN_usub: Always return non-negative results instead
1442 of not touching the result's sign bit.
1445 +) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
1449 +) Changed the LHASH code to use prototypes for callbacks, and created
1450 macros to declare and implement thin (optionally static) functions
1451 that provide type-safety and avoid function pointer casting for the
1452 type-specific callbacks.
1455 *) Use better test patterns in bntest.
1458 +) Added Kerberos Cipher Suites to be used with TLS, as written in
1460 [Veers Staats <staatsvr@asc.hpc.mil>,
1461 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
1463 *) rand_win.c fix for Borland C.
1466 *) BN_rshift bugfix for n == 0.
1469 +) Reformat the FAQ so the different questions and answers can be divided
1470 in sections depending on the subject.
1473 +) Have the zlib compression code load ZLIB.DLL dynamically under
1477 +) New function BN_mod_sqrt for computing square roots modulo a prime
1478 (using the probabilistic Tonelli-Shanks algorithm unless
1479 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
1480 be handled deterministically).
1481 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
1483 *) Add a 'bctest' script that checks for some known 'bc' bugs
1484 so that 'make test' does not abort just because 'bc' is broken.
1487 *) Store verify_result within SSL_SESSION also for client side to
1488 avoid potential security hole. (Re-used sessions on the client side
1489 always resulted in verify_result==X509_V_OK, not using the original
1490 result of the server certificate verification.)
1493 +) Make BN_mod_inverse faster by explicitly handling small quotients
1494 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
1495 512 bits], about 30% for larger ones [1024 or 2048 bits].)
1498 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
1499 SSL3_RT_APPLICATION_DATA, return 0.
1500 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
1504 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
1505 releases, have been re-implemented by renaming the previous
1506 implementations of ssl2_read and ssl3_read to ssl2_read_internal
1507 and ssl3_read_internal, respectively, and adding 'peek' parameters
1508 to them. The new ssl[23]_{read,peek} functions are calls to
1509 ssl[23]_read_internal with the 'peek' flag set appropriately.
1510 A 'peek' parameter has also been added to ssl3_read_bytes, which
1511 does the actual work for ssl3_read_internal.
1514 +) New function BN_kronecker.
1517 +) Fix BN_gcd so that it works on negative inputs; the result is
1518 positive unless both parameters are zero.
1519 Previously something reasonably close to an infinite loop was
1520 possible because numbers could be growing instead of shrinking
1521 in the implementation of Euclid's algorithm.
1524 +) Fix BN_is_word() and BN_is_one() macros to take into account the
1525 sign of the number in question.
1527 Fix BN_is_word(a,w) to work correctly for w == 0.
1529 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
1530 because its test if the absolute value of 'a' equals 'w'.
1531 Note that BN_abs_is_word does *not* handle w == 0 reliably;
1532 it exists mostly for use in the implementations of BN_is_zero(),
1533 BN_is_one(), and BN_is_word().
1536 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
1537 the method-specific "init()" handler. Also clean up ex_data after
1538 calling the method-specific "finish()" handler. Previously, this was
1539 happening the other way round.
1542 +) New function BN_swap.
1545 +) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
1546 the exponentiation functions are more likely to produce reasonable
1547 results on negative inputs.
1550 +) Change BN_mod_mul so that the result is always non-negative.
1551 Previously, it could be negative if one of the factors was negative;
1552 I don't think anyone really wanted that behaviour.
1555 +) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
1556 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
1557 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
1558 and add new functions:
1567 BN_mod_lshift1_quick
1571 These functions always generate non-negative results.
1573 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
1574 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
1576 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
1577 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
1578 be reduced modulo m.
1579 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
1581 +) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
1582 was actually never needed) and in BN_mul(). The removal in BN_mul()
1583 required a small change in bn_mul_part_recursive() and the addition
1584 of the functions bn_cmp_part_words(), bn_sub_part_words() and
1585 bn_add_part_words(), which do the same thing as bn_cmp_words(),
1586 bn_sub_words() and bn_add_words() except they take arrays with
1590 +) In 'openssl passwd', verify passwords read from the terminal
1591 unless the '-salt' option is used (which usually means that
1592 verification would just waste user's time since the resulting
1593 hash is going to be compared with some given password hash)
1594 or the new '-noverify' option is used.
1596 This is an incompatible change, but it does not affect
1597 non-interactive use of 'openssl passwd' (passwords on the command
1598 line, '-stdin' option, '-in ...' option) and thus should not
1602 +) Remove all references to RSAref, since there's no more need for it.
1605 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
1606 The previous value, 12, was not always sufficient for BN_mod_exp().
1609 +) Make DSO load along a path given through an environment variable
1610 (SHLIB_PATH) with shl_load().
1613 +) Constify the ENGINE code as a result of BIGNUM constification.
1614 Also constify the RSA code and most things related to it. In a
1615 few places, most notable in the depth of the ASN.1 code, ugly
1616 casts back to non-const were required (to be solved at a later
1620 +) Make it so the openssl application has all engines loaded by default.
1623 +) Constify the BIGNUM routines a little more.
1626 *) Make sure that shared libraries get the internal name engine with
1627 the full version number and not just 0. This should mark the
1628 shared libraries as not backward compatible. Of course, this should
1629 be changed again when we can guarantee backward binary compatibility.
1632 +) Add the following functions:
1634 ENGINE_load_cswift()
1636 ENGINE_load_atalla()
1638 ENGINE_load_builtin_engines()
1640 That way, an application can itself choose if external engines that
1641 are built-in in OpenSSL shall ever be used or not. The benefit is
1642 that applications won't have to be linked with libdl or other dso
1643 libraries unless it's really needed.
1645 Changed 'openssl engine' to load all engines on demand.
1646 Changed the engine header files to avoid the duplication of some
1647 declarations (they differed!).
1650 +) 'openssl engine' can now list capabilities.
1653 +) Better error reporting in 'openssl engine'.
1656 +) Never call load_dh_param(NULL) in s_server.
1659 +) Add engine application. It can currently list engines by name and
1660 identity, and test if they are actually available.
1663 +) Add support for shared libraries under Irix.
1664 [Albert Chin-A-Young <china@thewrittenword.com>]
1666 +) Improve RPM specification file by forcing symbolic linking and making
1667 sure the installed documentation is also owned by root.root.
1668 [Damien Miller <djm@mindrot.org>]
1670 +) Add configuration option to build on Linux on both big-endian and
1672 [Ralf Baechle <ralf@uni-koblenz.de>]
1674 +) Give the OpenSSL applications more possibilities to make use of
1675 keys (public as well as private) handled by engines.
1678 +) Add OCSP code that comes from CertCo.
1681 +) Add VMS support for the Rijndael code.
1684 +) Added untested support for Nuron crypto accelerator.
1687 +) Add support for external cryptographic devices. This code was
1688 previously distributed separately as the "engine" branch.
1689 [Geoff Thorpe, Richard Levitte]
1691 +) Rework the filename-translation in the DSO code. It is now possible to
1692 have far greater control over how a "name" is turned into a filename
1693 depending on the operating environment and any oddities about the
1694 different shared library filenames on each system.
1697 +) Support threads on FreeBSD-elf in Configure.
1700 +) Add the possibility to create shared libraries on HP-UX
1703 +) Fix for SHA1 assembly problem with MASM: it produces
1704 warnings about corrupt line number information when assembling
1705 with debugging information. This is caused by the overlapping
1707 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
1710 NCONF_get_number() has no error checking at all. As a replacement,
1711 NCONF_get_number_e() is defined (_e for "error checking") and is
1712 promoted strongly. The old NCONF_get_number is kept around for
1713 binary backward compatibility.
1714 Make it possible for methods to load from something other than a BIO,
1715 by providing a function pointer that is given a name instead of a BIO.
1716 For example, this could be used to load configuration data from an
1720 *) Fix typo in get_cert_by_subject() in by_dir.c
1721 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
1723 *) Rework the system to generate shared libraries:
1725 - Make note of the expected extension for the shared libraries and
1726 if there is a need for symbolic links from for example libcrypto.so.0
1727 to libcrypto.so.0.9.7. There is extended info in Configure for
1730 - Make as few rebuilds of the shared libraries as possible.
1732 - Still avoid linking the OpenSSL programs with the shared libraries.
1734 - When installing, install the shared libraries separately from the
1738 +) Fix for non blocking accept BIOs. Added new I/O special reason
1739 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
1740 with non blocking I/O was not possible because no retry code was
1741 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
1745 +) Added the beginnings of Rijndael support.
1748 +) Fix for bug in DirectoryString mask setting. Add support for
1749 X509_NAME_print_ex() in 'req' and X509_print_ex() function
1750 to allow certificate printing to more controllable, additional
1751 'certopt' option to 'x509' to allow new printing options to be
1755 +) Clean old EAY MD5 hack from e_os.h.
1758 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
1760 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
1761 and not in SSL_clear because the latter is also used by the
1762 accept/connect functions; previously, the settings made by
1763 SSL_set_read_ahead would be lost during the handshake.
1764 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
1766 *) Correct util/mkdef.pl to be selective about disabled algorithms.
1767 Previously, it would create entries for disableed algorithms no
1771 *) Added several new manual pages for SSL_* function.
1774 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
1776 *) In ssl23_get_client_hello, generate an error message when faced
1777 with an initial SSL 3.0/TLS record that is too small to contain the
1778 first two bytes of the ClientHello message, i.e. client_version.
1779 (Note that this is a pathologic case that probably has never happened
1780 in real life.) The previous approach was to use the version number
1781 from the record header as a substitute; but our protocol choice
1782 should not depend on that one because it is not authenticated
1783 by the Finished messages.
1786 *) More robust randomness gathering functions for Windows.
1787 [Jeffrey Altman <jaltman@columbia.edu>]
1789 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
1790 not set then we don't setup the error code for issuer check errors
1791 to avoid possibly overwriting other errors which the callback does
1792 handle. If an application does set the flag then we assume it knows
1793 what it is doing and can handle the new informational codes
1797 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
1798 a general "ANY" type, as such it should be able to decode anything
1799 including tagged types. However it didn't check the class so it would
1800 wrongly interpret tagged types in the same way as their universal
1801 counterpart and unknown types were just rejected. Changed so that the
1802 tagged and unknown types are handled in the same way as a SEQUENCE:
1803 that is the encoding is stored intact. There is also a new type
1804 "V_ASN1_OTHER" which is used when the class is not universal, in this
1805 case we have no idea what the actual type is so we just lump them all
1809 *) On VMS, stdout may very well lead to a file that is written to
1810 in a record-oriented fashion. That means that every write() will
1811 write a separate record, which will be read separately by the
1812 programs trying to read from it. This can be very confusing.
1814 The solution is to put a BIO filter in the way that will buffer
1815 text until a linefeed is reached, and then write everything a
1816 line at a time, so every record written will be an actual line,
1817 not chunks of lines and not (usually doesn't happen, but I've
1818 seen it once) several lines in one record. BIO_f_linebuffer() is
1821 Currently, it's a VMS-only method, because that's where it has
1822 been tested well enough.
1825 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
1826 it can return incorrect results.
1827 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
1828 but it was in 0.9.6-beta[12].)
1831 *) Disable the check for content being present when verifying detached
1832 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
1833 include zero length content when signing messages.
1836 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
1837 BIO_ctrl (for BIO pairs).
1840 *) Add DSO method for VMS.
1843 *) Bug fix: Montgomery multiplication could produce results with the
1847 *) Add RPM specification openssl.spec and modify it to build three
1848 packages. The default package contains applications, application
1849 documentation and run-time libraries. The devel package contains
1850 include files, static libraries and function documentation. The
1851 doc package contains the contents of the doc directory. The original
1852 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
1855 *) Add a large number of documentation files for many SSL routines.
1856 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
1858 *) Add a configuration entry for Sony News 4.
1859 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
1861 *) Don't set the two most significant bits to one when generating a
1862 random number < q in the DSA library.
1865 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
1866 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
1867 the underlying transport is blocking) if a handshake took place.
1868 (The default behaviour is needed by applications such as s_client
1869 and s_server that use select() to determine when to use SSL_read;
1870 but for applications that know in advance when to expect data, it
1871 just makes things more complicated.)
1874 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
1878 *) Add a few more EBCDIC conditionals that make `req' and `x509'
1879 work better on such systems.
1880 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
1882 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
1883 Update PKCS12_parse() so it copies the friendlyName and the
1884 keyid to the certificates aux info.
1887 *) Fix bug in PKCS7_verify() which caused an infinite loop
1888 if there was more than one signature.
1889 [Sven Uszpelkat <su@celocom.de>]
1891 *) Major change in util/mkdef.pl to include extra information
1892 about each symbol, as well as presentig variables as well
1893 as functions. This change means that there's n more need
1894 to rebuild the .num files when some algorithms are excluded.
1897 *) Allow the verify time to be set by an application,
1898 rather than always using the current time.
1901 *) Phase 2 verify code reorganisation. The certificate
1902 verify code now looks up an issuer certificate by a
1903 number of criteria: subject name, authority key id
1904 and key usage. It also verifies self signed certificates
1905 by the same criteria. The main comparison function is
1906 X509_check_issued() which performs these checks.
1908 Lot of changes were necessary in order to support this
1909 without completely rewriting the lookup code.
1911 Authority and subject key identifier are now cached.
1913 The LHASH 'certs' is X509_STORE has now been replaced
1914 by a STACK_OF(X509_OBJECT). This is mainly because an
1915 LHASH can't store or retrieve multiple objects with
1916 the same hash value.
1918 As a result various functions (which were all internal
1919 use only) have changed to handle the new X509_STORE
1920 structure. This will break anything that messed round
1921 with X509_STORE internally.
1923 The functions X509_STORE_add_cert() now checks for an
1924 exact match, rather than just subject name.
1926 The X509_STORE API doesn't directly support the retrieval
1927 of multiple certificates matching a given criteria, however
1928 this can be worked round by performing a lookup first
1929 (which will fill the cache with candidate certificates)
1930 and then examining the cache for matches. This is probably
1931 the best we can do without throwing out X509_LOOKUP
1932 entirely (maybe later...).
1934 The X509_VERIFY_CTX structure has been enhanced considerably.
1936 All certificate lookup operations now go via a get_issuer()
1937 callback. Although this currently uses an X509_STORE it
1938 can be replaced by custom lookups. This is a simple way
1939 to bypass the X509_STORE hackery necessary to make this
1940 work and makes it possible to use more efficient techniques
1941 in future. A very simple version which uses a simple
1942 STACK for its trusted certificate store is also provided
1943 using X509_STORE_CTX_trusted_stack().
1945 The verify_cb() and verify() callbacks now have equivalents
1946 in the X509_STORE_CTX structure.
1948 X509_STORE_CTX also has a 'flags' field which can be used
1949 to customise the verify behaviour.
1952 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
1953 excludes S/MIME capabilities.
1956 *) When a certificate request is read in keep a copy of the
1957 original encoding of the signed data and use it when outputing
1958 again. Signatures then use the original encoding rather than
1959 a decoded, encoded version which may cause problems if the
1960 request is improperly encoded.
1963 *) For consistency with other BIO_puts implementations, call
1964 buffer_write(b, ...) directly in buffer_puts instead of calling
1967 In BIO_puts, increment b->num_write as in BIO_write.
1968 [Peter.Sylvester@EdelWeb.fr]
1970 *) Fix BN_mul_word for the case where the word is 0. (We have to use
1971 BN_zero, we may not return a BIGNUM with an array consisting of
1975 *) Avoid calling abort() from within the library when problems are
1976 detected, except if preprocessor symbols have been defined
1977 (such as REF_CHECK, BN_DEBUG etc.).
1980 *) New openssl application 'rsautl'. This utility can be
1981 used for low level RSA operations. DER public key
1982 BIO/fp routines also added.
1985 *) New Configure entry and patches for compiling on QNX 4.
1986 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
1988 *) A demo state-machine implementation was sponsored by
1989 Nuron (http://www.nuron.com/) and is now available in
1990 demos/state_machine.
1993 *) New options added to the 'dgst' utility for signature
1994 generation and verification.
1997 *) Unrecognized PKCS#7 content types are now handled via a
1998 catch all ASN1_TYPE structure. This allows unsupported
1999 types to be stored as a "blob" and an application can
2000 encode and decode it manually.
2003 *) Fix various signed/unsigned issues to make a_strex.c
2005 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
2007 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
2008 length if passed a buffer. ASN1_INTEGER_to_BN failed
2009 if passed a NULL BN and its argument was negative.
2010 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
2012 *) Modification to PKCS#7 encoding routines to output definite
2013 length encoding. Since currently the whole structures are in
2014 memory there's not real point in using indefinite length
2015 constructed encoding. However if OpenSSL is compiled with
2016 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
2019 *) Added BIO_vprintf() and BIO_vsnprintf().
2022 *) Added more prefixes to parse for in the the strings written
2023 through a logging bio, to cover all the levels that are available
2024 through syslog. The prefixes are now:
2026 PANIC, EMERG, EMR => LOG_EMERG
2027 ALERT, ALR => LOG_ALERT
2028 CRIT, CRI => LOG_CRIT
2029 ERROR, ERR => LOG_ERR
2030 WARNING, WARN, WAR => LOG_WARNING
2031 NOTICE, NOTE, NOT => LOG_NOTICE
2032 INFO, INF => LOG_INFO
2033 DEBUG, DBG => LOG_DEBUG
2035 and as before, if none of those prefixes are present at the
2036 beginning of the string, LOG_ERR is chosen.
2038 On Win32, the LOG_* levels are mapped according to this:
2040 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
2041 LOG_WARNING => EVENTLOG_WARNING_TYPE
2042 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
2046 *) Made it possible to reconfigure with just the configuration
2047 argument "reconf" or "reconfigure". The command line arguments
2048 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
2049 and are retrieved from there when reconfiguring.
2053 [Assar Westerlund <assar@sics.se>, Richard Levitte]
2055 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
2058 *) The obj_dat.pl script was messing up the sorting of object
2059 names. The reason was that it compared the quoted version
2060 of strings as a result "OCSP" > "OCSP Signing" because
2061 " > SPACE. Changed script to store unquoted versions of
2062 names and add quotes on output. It was also omitting some
2063 names from the lookup table if they were given a default
2064 value (that is if SN is missing it is given the same
2065 value as LN and vice versa), these are now added on the
2066 grounds that if an object has a name we should be able to
2067 look it up. Finally added warning output when duplicate
2068 short or long names are found.
2071 *) Changes needed for Tandem NSK.
2072 [Scott Uroff <scott@xypro.com>]
2074 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
2075 RSA_padding_check_SSLv23(), special padding was never detected
2076 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
2077 version rollback attacks was not effective.
2079 In s23_clnt.c, don't use special rollback-attack detection padding
2080 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
2081 client; similarly, in s23_srvr.c, don't do the rollback check if
2082 SSL 2.0 is the only protocol enabled in the server.
2085 *) Make it possible to get hexdumps of unprintable data with 'openssl
2086 asn1parse'. By implication, the functions ASN1_parse_dump() and
2087 BIO_dump_indent() are added.
2090 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
2091 these print out strings and name structures based on various
2092 flags including RFC2253 support and proper handling of
2093 multibyte characters. Added options to the 'x509' utility
2094 to allow the various flags to be set.
2097 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
2098 Also change the functions X509_cmp_current_time() and
2099 X509_gmtime_adj() work with an ASN1_TIME structure,
2100 this will enable certificates using GeneralizedTime in validity
2101 dates to be checked.
2104 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
2105 negative public key encodings) on by default,
2106 NO_NEG_PUBKEY_BUG can be set to disable it.
2109 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
2110 content octets. An i2c_ASN1_OBJECT is unnecessary because
2111 the encoding can be trivially obtained from the structure.
2114 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
2115 not read locks (CRYPTO_r_[un]lock).
2118 *) A first attempt at creating official support for shared
2119 libraries through configuration. I've kept it so the
2120 default is static libraries only, and the OpenSSL programs
2121 are always statically linked for now, but there are
2122 preparations for dynamic linking in place.
2123 This has been tested on Linux and Tru64.
2126 *) Randomness polling function for Win9x, as described in:
2127 Peter Gutmann, Software Generation of Practically Strong
2131 *) Fix so PRNG is seeded in req if using an already existing
2135 *) New options to smime application. -inform and -outform
2136 allow alternative formats for the S/MIME message including
2137 PEM and DER. The -content option allows the content to be
2138 specified separately. This should allow things like Netscape
2139 form signing output easier to verify.
2142 *) Fix the ASN1 encoding of tags using the 'long form'.
2145 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
2146 STRING types. These convert content octets to and from the
2147 underlying type. The actual tag and length octets are
2148 already assumed to have been read in and checked. These
2149 are needed because all other string types have virtually
2150 identical handling apart from the tag. By having versions
2151 of the ASN1 functions that just operate on content octets
2152 IMPLICIT tagging can be handled properly. It also allows
2153 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
2154 and ASN1_INTEGER are identical apart from the tag.
2157 *) Change the handling of OID objects as follows:
2159 - New object identifiers are inserted in objects.txt, following
2160 the syntax given in objects.README.
2161 - objects.pl is used to process obj_mac.num and create a new
2163 - obj_dat.pl is used to create a new obj_dat.h, using the data in
2166 This is currently kind of a hack, and the perl code in objects.pl
2167 isn't very elegant, but it works as I intended. The simplest way
2168 to check that it worked correctly is to look in obj_dat.h and
2169 check the array nid_objs and make sure the objects haven't moved
2170 around (this is important!). Additions are OK, as well as
2171 consistent name changes.
2174 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
2177 *) Addition of the command line parameter '-rand file' to 'openssl req'.
2178 The given file adds to whatever has already been seeded into the
2179 random pool through the RANDFILE configuration file option or
2180 environment variable, or the default random state file.
2183 *) mkstack.pl now sorts each macro group into lexical order.
2184 Previously the output order depended on the order the files
2185 appeared in the directory, resulting in needless rewriting
2189 *) Patches to make OpenSSL compile under Win32 again. Mostly
2190 work arounds for the VC++ problem that it treats func() as
2191 func(void). Also stripped out the parts of mkdef.pl that
2192 added extra typesafe functions: these no longer exist.
2195 *) Reorganisation of the stack code. The macros are now all
2196 collected in safestack.h . Each macro is defined in terms of
2197 a "stack macro" of the form SKM_<name>(type, a, b). The
2198 DEBUG_SAFESTACK is now handled in terms of function casts,
2199 this has the advantage of retaining type safety without the
2200 use of additional functions. If DEBUG_SAFESTACK is not defined
2201 then the non typesafe macros are used instead. Also modified the
2202 mkstack.pl script to handle the new form. Needs testing to see
2203 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
2204 the default if no major problems. Similar behaviour for ASN1_SET_OF
2205 and PKCS12_STACK_OF.
2208 *) When some versions of IIS use the 'NET' form of private key the
2209 key derivation algorithm is different. Normally MD5(password) is
2210 used as a 128 bit RC4 key. In the modified case
2211 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
2212 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
2213 as the old Netscape_RSA functions except they have an additional
2214 'sgckey' parameter which uses the modified algorithm. Also added
2215 an -sgckey command line option to the rsa utility. Thanks to
2216 Adrian Peck <bertie@ncipher.com> for posting details of the modified
2217 algorithm to openssl-dev.
2220 *) The evp_local.h macros were using 'c.##kname' which resulted in
2221 invalid expansion on some systems (SCO 5.0.5 for example).
2222 Corrected to 'c.kname'.
2223 [Phillip Porch <root@theporch.com>]
2225 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
2226 a STACK of email addresses from a certificate or request, these look
2227 in the subject name and the subject alternative name extensions and
2228 omit any duplicate addresses.
2231 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
2232 This makes DSA verification about 2 % faster.
2235 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
2236 (meaning that now 2^5 values will be precomputed, which is only 4 KB
2237 plus overhead for 1024 bit moduli).
2238 This makes exponentiations about 0.5 % faster for 1024 bit
2239 exponents (as measured by "openssl speed rsa2048").
2242 *) Rename memory handling macros to avoid conflicts with other
2244 Malloc => OPENSSL_malloc
2245 Malloc_locked => OPENSSL_malloc_locked
2246 Realloc => OPENSSL_realloc
2247 Free => OPENSSL_free
2250 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
2251 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
2254 *) CygWin32 support.
2255 [John Jarvie <jjarvie@newsguy.com>]
2257 *) The type-safe stack code has been rejigged. It is now only compiled
2258 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
2259 by default all type-specific stack functions are "#define"d back to
2260 standard stack functions. This results in more streamlined output
2261 but retains the type-safety checking possibilities of the original
2265 *) The STACK code has been cleaned up, and certain type declarations
2266 that didn't make a lot of sense have been brought in line. This has
2267 also involved a cleanup of sorts in safestack.h to more correctly
2268 map type-safe stack functions onto their plain stack counterparts.
2269 This work has also resulted in a variety of "const"ifications of
2270 lots of the code, especially "_cmp" operations which should normally
2271 be prototyped with "const" parameters anyway.
2274 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
2275 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
2276 (The PRNG state consists of two parts, the large pool 'state' and 'md',
2277 where all of 'md' is used each time the PRNG is used, but 'state'
2278 is used only indexed by a cyclic counter. As entropy may not be
2279 well distributed from the beginning, 'md' is important as a
2280 chaining variable. However, the output function chains only half
2281 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
2282 all of 'md', and seeding with STATE_SIZE dummy bytes will result
2283 in all of 'state' being rewritten, with the new values depending
2284 on virtually all of 'md'. This overcomes the 80 bit limitation.)
2287 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
2288 the handshake is continued after ssl_verify_cert_chain();
2289 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
2290 can lead to 'unexplainable' connection aborts later.
2291 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
2293 *) Major EVP API cipher revision.
2294 Add hooks for extra EVP features. This allows various cipher
2295 parameters to be set in the EVP interface. Support added for variable
2296 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
2297 setting of RC2 and RC5 parameters.
2299 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
2302 Remove lots of duplicated code from the EVP library. For example *every*
2303 cipher init() function handles the 'iv' in the same way according to the
2304 cipher mode. They also all do nothing if the 'key' parameter is NULL and
2305 for CFB and OFB modes they zero ctx->num.
2307 New functionality allows removal of S/MIME code RC2 hack.
2309 Most of the routines have the same form and so can be declared in terms
2312 By shifting this to the top level EVP_CipherInit() it can be removed from
2313 all individual ciphers. If the cipher wants to handle IVs or keys
2314 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
2317 Change lots of functions like EVP_EncryptUpdate() to now return a
2318 value: although software versions of the algorithms cannot fail
2319 any installed hardware versions can.
2322 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
2323 this option is set, tolerate broken clients that send the negotiated
2324 protocol version number instead of the requested protocol version
2328 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
2329 i.e. non-zero for export ciphersuites, zero otherwise.
2330 Previous versions had this flag inverted, inconsistent with
2331 rsa_tmp_cb (..._TMP_RSA_CB).
2332 [Bodo Moeller; problem reported by Amit Chopra]
2334 *) Add missing DSA library text string. Work around for some IIS
2335 key files with invalid SEQUENCE encoding.
2338 *) Add a document (doc/standards.txt) that list all kinds of standards
2339 and so on that are implemented in OpenSSL.
2342 *) Enhance c_rehash script. Old version would mishandle certificates
2343 with the same subject name hash and wouldn't handle CRLs at all.
2344 Added -fingerprint option to crl utility, to support new c_rehash
2348 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
2351 *) Fix for SSL server purpose checking. Server checking was
2352 rejecting certificates which had extended key usage present
2353 but no ssl client purpose.
2354 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
2356 *) Make PKCS#12 code work with no password. The PKCS#12 spec
2357 is a little unclear about how a blank password is handled.
2358 Since the password in encoded as a BMPString with terminating
2359 double NULL a zero length password would end up as just the
2360 double NULL. However no password at all is different and is
2361 handled differently in the PKCS#12 key generation code. NS
2362 treats a blank password as zero length. MSIE treats it as no
2363 password on export: but it will try both on import. We now do
2364 the same: PKCS12_parse() tries zero length and no password if
2365 the password is set to "" or NULL (NULL is now a valid password:
2366 it wasn't before) as does the pkcs12 application.
2369 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
2370 perror when PEM_read_bio_X509_REQ fails, the error message must
2371 be obtained from the error queue.
2374 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
2375 it in ERR_remove_state if appropriate, and change ERR_get_state
2376 accordingly to avoid race conditions (this is necessary because
2377 thread_hash is no longer constant once set).
2380 *) Bugfix for linux-elf makefile.one.
2383 *) RSA_get_default_method() will now cause a default
2384 RSA_METHOD to be chosen if one doesn't exist already.
2385 Previously this was only set during a call to RSA_new()
2386 or RSA_new_method(NULL) meaning it was possible for
2387 RSA_get_default_method() to return NULL.
2390 *) Added native name translation to the existing DSO code
2391 that will convert (if the flag to do so is set) filenames
2392 that are sufficiently small and have no path information
2393 into a canonical native form. Eg. "blah" converted to
2394 "libblah.so" or "blah.dll" etc.
2397 *) New function ERR_error_string_n(e, buf, len) which is like
2398 ERR_error_string(e, buf), but writes at most 'len' bytes
2399 including the 0 terminator. For ERR_error_string_n, 'buf'
2401 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
2403 *) CONF library reworked to become more general. A new CONF
2404 configuration file reader "class" is implemented as well as a
2405 new functions (NCONF_*, for "New CONF") to handle it. The now
2406 old CONF_* functions are still there, but are reimplemented to
2407 work in terms of the new functions. Also, a set of functions
2408 to handle the internal storage of the configuration data is
2409 provided to make it easier to write new configuration file
2410 reader "classes" (I can definitely see something reading a
2411 configuration file in XML format, for example), called _CONF_*,
2412 or "the configuration storage API"...
2414 The new configuration file reading functions are:
2416 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
2417 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
2419 NCONF_default, NCONF_WIN32
2421 NCONF_dump_fp, NCONF_dump_bio
2423 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
2424 NCONF_new creates a new CONF object. This works in the same way
2425 as other interfaces in OpenSSL, like the BIO interface.
2426 NCONF_dump_* dump the internal storage of the configuration file,
2427 which is useful for debugging. All other functions take the same
2428 arguments as the old CONF_* functions wth the exception of the
2429 first that must be a `CONF *' instead of a `LHASH *'.
2431 To make it easer to use the new classes with the old CONF_* functions,
2432 the function CONF_set_default_method is provided.
2435 *) Add '-tls1' option to 'openssl ciphers', which was already
2436 mentioned in the documentation but had not been implemented.
2437 (This option is not yet really useful because even the additional
2438 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
2441 *) Initial DSO code added into libcrypto for letting OpenSSL (and
2442 OpenSSL-based applications) load shared libraries and bind to
2443 them in a portable way.
2444 [Geoff Thorpe, with contributions from Richard Levitte]
2446 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
2448 *) Make sure _lrotl and _lrotr are only used with MSVC.
2450 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
2451 (the default implementation of RAND_status).
2453 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
2454 to '-clrext' (= clear extensions), as intended and documented.
2455 [Bodo Moeller; inconsistency pointed out by Michael Attili
2458 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
2459 was larger than the MD block size.
2460 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
2462 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
2463 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
2464 using the passed key: if the passed key was a private key the result
2465 of X509_print(), for example, would be to print out all the private key
2469 *) des_quad_cksum() byte order bug fix.
2470 [Ulf Möller, using the problem description in krb4-0.9.7, where
2471 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
2473 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
2475 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
2477 *) For easily testing in shell scripts whether some command
2478 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
2479 returns with exit code 0 iff no command of the given name is available.
2480 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
2481 the output goes to stdout and nothing is printed to stderr.
2482 Additional arguments are always ignored.
2484 Since for each cipher there is a command of the same name,
2485 the 'no-cipher' compilation switches can be tested this way.
2487 ('openssl no-XXX' is not able to detect pseudo-commands such
2488 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
2491 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
2494 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
2495 is set; it will be thrown away anyway because each handshake creates
2497 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
2498 to parameters -- in previous versions (since OpenSSL 0.9.3) the
2499 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
2500 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
2503 *) New s_client option -ign_eof: EOF at stdin is ignored, and
2504 'Q' and 'R' lose their special meanings (quit/renegotiate).
2505 This is part of what -quiet does; unlike -quiet, -ign_eof
2506 does not suppress any output.
2509 *) Add compatibility options to the purpose and trust code. The
2510 purpose X509_PURPOSE_ANY is "any purpose" which automatically
2511 accepts a certificate or CA, this was the previous behaviour,
2512 with all the associated security issues.
2514 X509_TRUST_COMPAT is the old trust behaviour: only and
2515 automatically trust self signed roots in certificate store. A
2516 new trust setting X509_TRUST_DEFAULT is used to specify that
2517 a purpose has no associated trust setting and it should instead
2518 use the value in the default purpose.
2521 *) Fix the PKCS#8 DSA private key code so it decodes keys again
2522 and fix a memory leak.
2525 *) In util/mkerr.pl (which implements 'make errors'), preserve
2526 reason strings from the previous version of the .c file, as
2527 the default to have only downcase letters (and digits) in
2528 automatically generated reasons codes is not always appropriate.
2531 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
2532 using strerror. Previously, ERR_reason_error_string() returned
2533 library names as reason strings for SYSerr; but SYSerr is a special
2534 case where small numbers are errno values, not library numbers.
2537 *) Add '-dsaparam' option to 'openssl dhparam' application. This
2538 converts DSA parameters into DH parameters. (When creating parameters,
2539 DSA_generate_parameters is used.)
2542 *) Include 'length' (recommended exponent length) in C code generated
2543 by 'openssl dhparam -C'.
2546 *) The second argument to set_label in perlasm was already being used
2547 so couldn't be used as a "file scope" flag. Moved to third argument
2551 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
2552 instead of RAND_bytes for encryption IVs and salts.
2555 *) Include RAND_status() into RAND_METHOD instead of implementing
2556 it only for md_rand.c Otherwise replacing the PRNG by calling
2557 RAND_set_rand_method would be impossible.
2560 *) Don't let DSA_generate_key() enter an infinite loop if the random
2561 number generation fails.
2564 *) New 'rand' application for creating pseudo-random output.
2567 *) Added configuration support for Linux/IA64
2568 [Rolf Haberrecker <rolf@suse.de>]
2570 *) Assembler module support for Mingw32.
2573 *) Shared library support for HPUX (in shlib/).
2574 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
2576 *) Shared library support for Solaris gcc.
2577 [Lutz Behnke <behnke@trustcenter.de>]
2579 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
2581 *) PKCS7_encrypt() was adding text MIME headers twice because they
2582 were added manually and by SMIME_crlf_copy().
2585 *) In bntest.c don't call BN_rand with zero bits argument.
2586 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
2588 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
2589 case was implemented. This caused BN_div_recp() to fail occasionally.
2592 *) Add an optional second argument to the set_label() in the perl
2593 assembly language builder. If this argument exists and is set
2594 to 1 it signals that the assembler should use a symbol whose
2595 scope is the entire file, not just the current function. This
2596 is needed with MASM which uses the format label:: for this scope.
2597 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
2599 *) Change the ASN1 types so they are typedefs by default. Before
2600 almost all types were #define'd to ASN1_STRING which was causing
2601 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
2605 *) Change names of new functions to the new get1/get0 naming
2606 convention: After 'get1', the caller owns a reference count
2607 and has to call ..._free; 'get0' returns a pointer to some
2608 data structure without incrementing reference counters.
2609 (Some of the existing 'get' functions increment a reference
2610 counter, some don't.)
2611 Similarly, 'set1' and 'add1' functions increase reference
2612 counters or duplicate objects.
2615 *) Allow for the possibility of temp RSA key generation failure:
2616 the code used to assume it always worked and crashed on failure.
2619 *) Fix potential buffer overrun problem in BIO_printf().
2620 [Ulf Möller, using public domain code by Patrick Powell; problem
2621 pointed out by David Sacerdote <das33@cornell.edu>]
2623 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
2624 RAND_egd() and RAND_status(). In the command line application,
2625 the EGD socket can be specified like a seed file using RANDFILE
2629 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
2630 Some CAs (e.g. Verisign) distribute certificates in this form.
2633 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
2634 list to exclude them. This means that no special compilation option
2635 is needed to use anonymous DH: it just needs to be included in the
2639 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
2640 EVP_MD_type. The old functionality is available in a new macro called
2641 EVP_MD_md(). Change code that uses it and update docs.
2644 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
2645 where the 'void *' argument is replaced by a function pointer argument.
2646 Previously 'void *' was abused to point to functions, which works on
2647 many platforms, but is not correct. As these functions are usually
2648 called by macros defined in OpenSSL header files, most source code
2649 should work without changes.
2652 *) <openssl/opensslconf.h> (which is created by Configure) now contains
2653 sections with information on -D... compiler switches used for
2654 compiling the library so that applications can see them. To enable
2655 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
2656 must be defined. E.g.,
2657 #define OPENSSL_ALGORITHM_DEFINES
2658 #include <openssl/opensslconf.h>
2659 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
2660 [Richard Levitte, Ulf and Bodo Möller]
2662 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
2666 *) Change the 'other' type in certificate aux info to a STACK_OF
2667 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
2668 the required ASN1 format: arbitrary types determined by an OID.
2671 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
2672 argument to 'req'. This is not because the function is newer or
2673 better than others it just uses the work 'NEW' in the certificate
2674 request header lines. Some software needs this.
2677 *) Reorganise password command line arguments: now passwords can be
2678 obtained from various sources. Delete the PEM_cb function and make
2679 it the default behaviour: i.e. if the callback is NULL and the
2680 usrdata argument is not NULL interpret it as a null terminated pass
2681 phrase. If usrdata and the callback are NULL then the pass phrase
2682 is prompted for as usual.
2685 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
2686 the support is automatically enabled. The resulting binaries will
2687 autodetect the card and use it if present.
2688 [Ben Laurie and Compaq Inc.]
2690 *) Work around for Netscape hang bug. This sends certificate request
2691 and server done in one record. Since this is perfectly legal in the
2692 SSL/TLS protocol it isn't a "bug" option and is on by default. See
2693 the bugs/SSLv3 entry for more info.
2696 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
2699 *) Add -rand argument to smime and pkcs12 applications and read/write
2703 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
2706 *) Add command line password options to the remaining applications.
2709 *) Bug fix for BN_div_recp() for numerators with an even number of
2713 *) More tests in bntest.c, and changed test_bn output.
2716 *) ./config recognizes MacOS X now.
2719 *) Bug fix for BN_div() when the first words of num and divsor are
2720 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
2723 *) Add support for various broken PKCS#8 formats, and command line
2724 options to produce them.
2727 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
2728 get temporary BIGNUMs from a BN_CTX.
2731 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
2735 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
2736 include a #define from the old name to the new. The original intent
2737 was that statically linked binaries could for example just call
2738 SSLeay_add_all_ciphers() to just add ciphers to the table and not
2739 link with digests. This never worked becayse SSLeay_add_all_digests()
2740 and SSLeay_add_all_ciphers() were in the same source file so calling
2741 one would link with the other. They are now in separate source files.
2744 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
2747 *) Use a less unusual form of the Miller-Rabin primality test (it used
2748 a binary algorithm for exponentiation integrated into the Miller-Rabin
2749 loop, our standard modexp algorithms are faster).
2752 *) Support for the EBCDIC character set completed.
2753 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
2755 *) Source code cleanups: use const where appropriate, eliminate casts,
2756 use void * instead of char * in lhash.
2759 *) Bugfix: ssl3_send_server_key_exchange was not restartable
2760 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
2761 this the server could overwrite ephemeral keys that the client
2765 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
2766 using 50 iterations of the Rabin-Miller test.
2768 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
2769 iterations of the Rabin-Miller test as required by the appendix
2770 to FIPS PUB 186[-1]) instead of DSA_is_prime.
2771 As BN_is_prime_fasttest includes trial division, DSA parameter
2772 generation becomes much faster.
2774 This implies a change for the callback functions in DSA_is_prime
2775 and DSA_generate_parameters: The callback function is called once
2776 for each positive witness in the Rabin-Miller test, not just
2777 occasionally in the inner loop; and the parameters to the
2778 callback function now provide an iteration count for the outer
2779 loop rather than for the current invocation of the inner loop.
2780 DSA_generate_parameters additionally can call the callback
2781 function with an 'iteration count' of -1, meaning that a
2782 candidate has passed the trial division test (when q is generated
2783 from an application-provided seed, trial division is skipped).
2786 *) New function BN_is_prime_fasttest that optionally does trial
2787 division before starting the Rabin-Miller test and has
2788 an additional BN_CTX * argument (whereas BN_is_prime always
2789 has to allocate at least one BN_CTX).
2790 'callback(1, -1, cb_arg)' is called when a number has passed the
2791 trial division stage.
2794 *) Fix for bug in CRL encoding. The validity dates weren't being handled
2798 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
2801 *) New function BN_pseudo_rand().
2804 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
2805 bignum version of BN_from_montgomery() with the working code from
2806 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
2810 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
2811 made it impossible to use the same SSL_SESSION data structure in
2812 SSL2 clients in multiple threads.
2815 *) The return value of RAND_load_file() no longer counts bytes obtained
2816 by stat(). RAND_load_file(..., -1) is new and uses the complete file
2817 to seed the PRNG (previously an explicit byte count was required).
2818 [Ulf Möller, Bodo Möller]
2820 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
2821 used (char *) instead of (void *) and had casts all over the place.
2824 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
2827 *) Retain source code compatibility for BN_prime_checks macro:
2828 BN_is_prime(..., BN_prime_checks, ...) now uses
2829 BN_prime_checks_for_size to determine the appropriate number of
2830 Rabin-Miller iterations.
2833 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
2834 DH_CHECK_P_NOT_SAFE_PRIME.
2835 (Check if this is true? OpenPGP calls them "strong".)
2838 *) Merge the functionality of "dh" and "gendh" programs into a new program
2839 "dhparam". The old programs are retained for now but will handle DH keys
2840 (instead of parameters) in future.
2843 *) Make the ciphers, s_server and s_client programs check the return values
2844 when a new cipher list is set.
2847 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
2848 ciphers. Before when the 56bit ciphers were enabled the sorting was
2851 The syntax for the cipher sorting has been extended to support sorting by
2852 cipher-strength (using the strength_bits hard coded in the tables).
2853 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
2855 Fix a bug in the cipher-command parser: when supplying a cipher command
2856 string with an "undefined" symbol (neither command nor alphanumeric
2857 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
2858 an error is flagged.
2860 Due to the strength-sorting extension, the code of the
2861 ssl_create_cipher_list() function was completely rearranged. I hope that
2862 the readability was also increased :-)
2863 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
2865 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
2866 for the first serial number and places 2 in the serial number file. This
2867 avoids problems when the root CA is created with serial number zero and
2868 the first user certificate has the same issuer name and serial number
2872 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
2873 the new code. Add documentation for this stuff.
2876 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
2877 X509_*() to X509at_*() on the grounds that they don't handle X509
2878 structures and behave in an analagous way to the X509v3 functions:
2879 they shouldn't be called directly but wrapper functions should be used
2882 So we also now have some wrapper functions that call the X509at functions
2883 when passed certificate requests. (TO DO: similar things can be done with
2884 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
2885 things. Some of these need some d2i or i2d and print functionality
2886 because they handle more complex structures.)
2889 *) Add missing #ifndefs that caused missing symbols when building libssl
2890 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
2891 NO_RSA in ssl/s2*.c.
2892 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
2894 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
2895 has a return value which indicates the quality of the random data
2896 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
2897 error queue. New function RAND_pseudo_bytes() generates output that is
2898 guaranteed to be unique but not unpredictable. RAND_add is like
2899 RAND_seed, but takes an extra argument for an entropy estimate
2900 (RAND_seed always assumes full entropy).
2903 *) Do more iterations of Rabin-Miller probable prime test (specifically,
2904 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
2905 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
2906 in crypto/bn/bn_prime.c for the complete table). This guarantees a
2907 false-positive rate of at most 2^-80 for random input.
2910 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
2913 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
2914 in the 0.9.5 release), this returns the chain
2915 from an X509_CTX structure with a dup of the stack and all
2916 the X509 reference counts upped: so the stack will exist
2917 after X509_CTX_cleanup() has been called. Modify pkcs12.c
2920 Also make SSL_SESSION_print() print out the verify return
2924 *) Add manpage for the pkcs12 command. Also change the default
2925 behaviour so MAC iteration counts are used unless the new
2926 -nomaciter option is used. This improves file security and
2927 only older versions of MSIE (4.0 for example) need it.
2930 *) Honor the no-xxx Configure options when creating .DEF files.
2933 *) Add PKCS#10 attributes to field table: challengePassword,
2934 unstructuredName and unstructuredAddress. These are taken from
2935 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
2936 international characters are used.
2938 More changes to X509_ATTRIBUTE code: allow the setting of types
2939 based on strings. Remove the 'loc' parameter when adding
2940 attributes because these will be a SET OF encoding which is sorted
2944 *) Initial changes to the 'req' utility to allow request generation
2945 automation. This will allow an application to just generate a template
2946 file containing all the field values and have req construct the
2949 Initial support for X509_ATTRIBUTE handling. Stacks of these are
2950 used all over the place including certificate requests and PKCS#7
2951 structures. They are currently handled manually where necessary with
2952 some primitive wrappers for PKCS#7. The new functions behave in a
2953 manner analogous to the X509 extension functions: they allow
2954 attributes to be looked up by NID and added.
2956 Later something similar to the X509V3 code would be desirable to
2957 automatically handle the encoding, decoding and printing of the
2958 more complex types. The string types like challengePassword can
2959 be handled by the string table functions.
2961 Also modified the multi byte string table handling. Now there is
2962 a 'global mask' which masks out certain types. The table itself
2963 can use the flag STABLE_NO_MASK to ignore the mask setting: this
2964 is useful when for example there is only one permissible type
2965 (as in countryName) and using the mask might result in no valid
2969 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
2970 SSL_get_peer_finished to allow applications to obtain the latest
2971 Finished messages sent to the peer or expected from the peer,
2972 respectively. (SSL_get_peer_finished is usually the Finished message
2973 actually received from the peer, otherwise the protocol will be aborted.)
2975 As the Finished message are message digests of the complete handshake
2976 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
2977 be used for external authentication procedures when the authentication
2978 provided by SSL/TLS is not desired or is not enough.
2981 *) Enhanced support for Alpha Linux is added. Now ./config checks if
2982 the host supports BWX extension and if Compaq C is present on the
2983 $PATH. Just exploiting of the BWX extension results in 20-30%
2984 performance kick for some algorithms, e.g. DES and RC4 to mention
2985 a couple. Compaq C in turn generates ~20% faster code for MD5 and
2989 *) Add support for MS "fast SGC". This is arguably a violation of the
2990 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
2991 weak crypto and after checking the certificate is SGC a second one
2992 with strong crypto. MS SGC stops the first handshake after receiving
2993 the server certificate message and sends a second client hello. Since
2994 a server will typically do all the time consuming operations before
2995 expecting any further messages from the client (server key exchange
2996 is the most expensive) there is little difference between the two.
2998 To get OpenSSL to support MS SGC we have to permit a second client
2999 hello message after we have sent server done. In addition we have to
3000 reset the MAC if we do get this second client hello.
3003 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
3004 if a DER encoded private key is RSA or DSA traditional format. Changed
3005 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
3006 format DER encoded private key. Newer code should use PKCS#8 format which
3007 has the key type encoded in the ASN1 structure. Added DER private key
3008 support to pkcs8 application.
3011 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
3012 ciphersuites has been selected (as required by the SSL 3/TLS 1
3013 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
3014 is set, we interpret this as a request to violate the specification
3015 (the worst that can happen is a handshake failure, and 'correct'
3016 behaviour would result in a handshake failure anyway).
3019 *) In SSL_CTX_add_session, take into account that there might be multiple
3020 SSL_SESSION structures with the same session ID (e.g. when two threads
3021 concurrently obtain them from an external cache).
3022 The internal cache can handle only one SSL_SESSION with a given ID,
3023 so if there's a conflict, we now throw out the old one to achieve
3027 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
3028 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
3029 some routines that use cipher OIDs: some ciphers do not have OIDs
3030 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
3034 *) Simplify the trust setting structure and code. Now we just have
3035 two sequences of OIDs for trusted and rejected settings. These will
3036 typically have values the same as the extended key usage extension
3037 and any application specific purposes.
3039 The trust checking code now has a default behaviour: it will just
3040 check for an object with the same NID as the passed id. Functions can
3041 be provided to override either the default behaviour or the behaviour
3042 for a given id. SSL client, server and email already have functions
3043 in place for compatibility: they check the NID and also return "trusted"
3044 if the certificate is self signed.
3047 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
3048 traditional format into an EVP_PKEY structure.
3051 *) Add a password callback function PEM_cb() which either prompts for
3052 a password if usr_data is NULL or otherwise assumes it is a null
3053 terminated password. Allow passwords to be passed on command line
3054 environment or config files in a few more utilities.
3057 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
3058 keys. Add some short names for PKCS#8 PBE algorithms and allow them
3059 to be specified on the command line for the pkcs8 and pkcs12 utilities.
3060 Update documentation.
3063 *) Support for ASN1 "NULL" type. This could be handled before by using
3064 ASN1_TYPE but there wasn't any function that would try to read a NULL
3065 and produce an error if it couldn't. For compatibility we also have
3066 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
3067 don't allocate anything because they don't need to.
3070 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
3072 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
3074 *) Rebuild of the memory allocation routines used by OpenSSL code and
3075 possibly others as well. The purpose is to make an interface that
3076 provide hooks so anyone can build a separate set of allocation and
3077 deallocation routines to be used by OpenSSL, for example memory
3078 pool implementations, or something else, which was previously hard
3079 since Malloc(), Realloc() and Free() were defined as macros having
3080 the values malloc, realloc and free, respectively (except for Win32
3081 compilations). The same is provided for memory debugging code.
3082 OpenSSL already comes with functionality to find memory leaks, but
3083 this gives people a chance to debug other memory problems.
3085 With these changes, a new set of functions and macros have appeared:
3087 CRYPTO_set_mem_debug_functions() [F]
3088 CRYPTO_get_mem_debug_functions() [F]
3089 CRYPTO_dbg_set_options() [F]
3090 CRYPTO_dbg_get_options() [F]
3091 CRYPTO_malloc_debug_init() [M]
3093 The memory debug functions are NULL by default, unless the library
3094 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
3095 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
3096 gives the standard debugging functions that come with OpenSSL) or
3097 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
3098 provided by the library user) must be used. When the standard
3099 debugging functions are used, CRYPTO_dbg_set_options can be used to
3100 request additional information:
3101 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
3102 the CRYPTO_MDEBUG_xxx macro when compiling the library.
3104 Also, things like CRYPTO_set_mem_functions will always give the
3105 expected result (the new set of functions is used for allocation
3106 and deallocation) at all times, regardless of platform and compiler
3109 To finish it up, some functions that were never use in any other
3110 way than through macros have a new API and new semantic:
3113 CRYPTO_dbg_realloc()
3116 All macros of value have retained their old syntax.
3117 [Richard Levitte and Bodo Moeller]
3119 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
3120 ordering of SMIMECapabilities wasn't in "strength order" and there
3121 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
3125 *) Some ASN1 types with illegal zero length encoding (INTEGER,
3126 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
3127 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
3129 *) Merge in my S/MIME library for OpenSSL. This provides a simple
3130 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
3131 functionality to handle multipart/signed properly) and a utility
3132 called 'smime' to call all this stuff. This is based on code I
3133 originally wrote for Celo who have kindly allowed it to be
3134 included in OpenSSL.
3137 *) Add variants des_set_key_checked and des_set_key_unchecked of
3138 des_set_key (aka des_key_sched). Global variable des_check_key
3139 decides which of these is called by des_set_key; this way
3140 des_check_key behaves as it always did, but applications and
3141 the library itself, which was buggy for des_check_key == 1,
3142 have a cleaner way to pick the version they need.
3145 *) New function PKCS12_newpass() which changes the password of a
3149 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
3150 dynamic mix. In both cases the ids can be used as an index into the
3151 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
3152 functions so they accept a list of the field values and the
3153 application doesn't need to directly manipulate the X509_TRUST
3157 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
3161 *) Modify the way the V3 extension code looks up extensions. This now
3162 works in a similar way to the object code: we have some "standard"
3163 extensions in a static table which is searched with OBJ_bsearch()
3164 and the application can add dynamic ones if needed. The file
3165 crypto/x509v3/ext_dat.h now has the info: this file needs to be
3166 updated whenever a new extension is added to the core code and kept
3167 in ext_nid order. There is a simple program 'tabtest.c' which checks
3168 this. New extensions are not added too often so this file can readily
3169 be maintained manually.
3171 There are two big advantages in doing things this way. The extensions
3172 can be looked up immediately and no longer need to be "added" using
3173 X509V3_add_standard_extensions(): this function now does nothing.
3174 [Side note: I get *lots* of email saying the extension code doesn't
3175 work because people forget to call this function]
3176 Also no dynamic allocation is done unless new extensions are added:
3177 so if we don't add custom extensions there is no need to call
3178 X509V3_EXT_cleanup().
3181 *) Modify enc utility's salting as follows: make salting the default. Add a
3182 magic header, so unsalted files fail gracefully instead of just decrypting
3183 to garbage. This is because not salting is a big security hole, so people
3184 should be discouraged from doing it.
3187 *) Fixes and enhancements to the 'x509' utility. It allowed a message
3188 digest to be passed on the command line but it only used this
3189 parameter when signing a certificate. Modified so all relevant
3190 operations are affected by the digest parameter including the
3191 -fingerprint and -x509toreq options. Also -x509toreq choked if a
3192 DSA key was used because it didn't fix the digest.
3195 *) Initial certificate chain verify code. Currently tests the untrusted
3196 certificates for consistency with the verify purpose (which is set
3197 when the X509_STORE_CTX structure is set up) and checks the pathlength.
3199 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
3200 this is because it will reject chains with invalid extensions whereas
3201 every previous version of OpenSSL and SSLeay made no checks at all.
3203 Trust code: checks the root CA for the relevant trust settings. Trust
3204 settings have an initial value consistent with the verify purpose: e.g.
3205 if the verify purpose is for SSL client use it expects the CA to be
3206 trusted for SSL client use. However the default value can be changed to
3207 permit custom trust settings: one example of this would be to only trust
3208 certificates from a specific "secure" set of CAs.
3210 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
3211 which should be used for version portability: especially since the
3212 verify structure is likely to change more often now.
3214 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
3215 to set them. If not set then assume SSL clients will verify SSL servers
3218 Two new options to the verify program: -untrusted allows a set of
3219 untrusted certificates to be passed in and -purpose which sets the
3220 intended purpose of the certificate. If a purpose is set then the
3221 new chain verify code is used to check extension consistency.
3224 *) Support for the authority information access extension.
3227 *) Modify RSA and DSA PEM read routines to transparently handle
3228 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
3229 public keys in a format compatible with certificate
3230 SubjectPublicKeyInfo structures. Unfortunately there were already
3231 functions called *_PublicKey_* which used various odd formats so
3232 these are retained for compatibility: however the DSA variants were
3233 never in a public release so they have been deleted. Changed dsa/rsa
3234 utilities to handle the new format: note no releases ever handled public
3235 keys so we should be OK.
3237 The primary motivation for this change is to avoid the same fiasco
3238 that dogs private keys: there are several incompatible private key
3239 formats some of which are standard and some OpenSSL specific and
3240 require various evil hacks to allow partial transparent handling and
3241 even then it doesn't work with DER formats. Given the option anything
3242 other than PKCS#8 should be dumped: but the other formats have to
3243 stay in the name of compatibility.
3245 With public keys and the benefit of hindsight one standard format
3246 is used which works with EVP_PKEY, RSA or DSA structures: though
3247 it clearly returns an error if you try to read the wrong kind of key.
3249 Added a -pubkey option to the 'x509' utility to output the public key.
3250 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
3251 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
3252 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
3253 that do the same as the EVP_PKEY_assign_*() except they up the
3254 reference count of the added key (they don't "swallow" the
3258 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
3259 CRLs would fail if the file contained no certificates or no CRLs:
3260 added a new function to read in both types and return the number
3261 read: this means that if none are read it will be an error. The
3262 DER versions of the certificate and CRL reader would always fail
3263 because it isn't possible to mix certificates and CRLs in DER format
3264 without choking one or the other routine. Changed this to just read
3265 a certificate: this is the best we can do. Also modified the code
3266 in apps/verify.c to take notice of return codes: it was previously
3267 attempting to read in certificates from NULL pointers and ignoring
3268 any errors: this is one reason why the cert and CRL reader seemed
3269 to work. It doesn't check return codes from the default certificate
3270 routines: these may well fail if the certificates aren't installed.
3273 *) Code to support otherName option in GeneralName.
3276 *) First update to verify code. Change the verify utility
3277 so it warns if it is passed a self signed certificate:
3278 for consistency with the normal behaviour. X509_verify
3279 has been modified to it will now verify a self signed
3280 certificate if *exactly* the same certificate appears
3281 in the store: it was previously impossible to trust a
3282 single self signed certificate. This means that:
3283 openssl verify ss.pem
3284 now gives a warning about a self signed certificate but
3285 openssl verify -CAfile ss.pem ss.pem
3289 *) For servers, store verify_result in SSL_SESSION data structure
3290 (and add it to external session representation).
3291 This is needed when client certificate verifications fails,
3292 but an application-provided verification callback (set by
3293 SSL_CTX_set_cert_verify_callback) allows accepting the session
3294 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
3295 but returns 1): When the session is reused, we have to set
3296 ssl->verify_result to the appropriate error code to avoid
3298 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
3300 *) Fix a bug in the new PKCS#7 code: it didn't consider the
3301 case in PKCS7_dataInit() where the signed PKCS7 structure
3302 didn't contain any existing data because it was being created.
3303 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
3305 *) Add a salt to the key derivation routines in enc.c. This
3306 forms the first 8 bytes of the encrypted file. Also add a
3307 -S option to allow a salt to be input on the command line.
3310 *) New function X509_cmp(). Oddly enough there wasn't a function
3311 to compare two certificates. We do this by working out the SHA1
3312 hash and comparing that. X509_cmp() will be needed by the trust
3316 *) SSL_get1_session() is like SSL_get_session(), but increments
3317 the reference count in the SSL_SESSION returned.
3318 [Geoff Thorpe <geoff@eu.c2.net>]
3320 *) Fix for 'req': it was adding a null to request attributes.
3321 Also change the X509_LOOKUP and X509_INFO code to handle
3322 certificate auxiliary information.
3325 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
3329 *) Add the possibility to add extra information to the memory leak
3330 detecting output, to form tracebacks, showing from where each
3331 allocation was originated: CRYPTO_push_info("constant string") adds
3332 the string plus current file name and line number to a per-thread
3333 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
3334 is like calling CYRPTO_pop_info() until the stack is empty.
3335 Also updated memory leak detection code to be multi-thread-safe.
3338 *) Add options -text and -noout to pkcs7 utility and delete the
3339 encryption options which never did anything. Update docs.
3342 *) Add options to some of the utilities to allow the pass phrase
3343 to be included on either the command line (not recommended on
3344 OSes like Unix) or read from the environment. Update the
3345 manpages and fix a few bugs.
3348 *) Add a few manpages for some of the openssl commands.
3351 *) Fix the -revoke option in ca. It was freeing up memory twice,
3352 leaking and not finding already revoked certificates.
3355 *) Extensive changes to support certificate auxiliary information.
3356 This involves the use of X509_CERT_AUX structure and X509_AUX
3357 functions. An X509_AUX function such as PEM_read_X509_AUX()
3358 can still read in a certificate file in the usual way but it
3359 will also read in any additional "auxiliary information". By
3360 doing things this way a fair degree of compatibility can be
3361 retained: existing certificates can have this information added
3362 using the new 'x509' options.
3364 Current auxiliary information includes an "alias" and some trust
3365 settings. The trust settings will ultimately be used in enhanced
3366 certificate chain verification routines: currently a certificate
3367 can only be trusted if it is self signed and then it is trusted
3371 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
3372 The problem was that one of the replacement routines had not been working
3373 since SSLeay releases. For now the offending routine has been replaced
3374 with non-optimised assembler. Even so, this now gives around 95%
3375 performance improvement for 1024 bit RSA signs.
3378 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
3379 handling. Most clients have the effective key size in bits equal to
3380 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
3381 A few however don't do this and instead use the size of the decrypted key
3382 to determine the RC2 key length and the AlgorithmIdentifier to determine
3383 the effective key length. In this case the effective key length can still
3384 be 40 bits but the key length can be 168 bits for example. This is fixed
3385 by manually forcing an RC2 key into the EVP_PKEY structure because the
3386 EVP code can't currently handle unusual RC2 key sizes: it always assumes
3387 the key length and effective key length are equal.
3390 *) Add a bunch of functions that should simplify the creation of
3391 X509_NAME structures. Now you should be able to do:
3392 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
3393 and have it automatically work out the correct field type and fill in
3394 the structures. The more adventurous can try:
3395 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
3396 and it will (hopefully) work out the correct multibyte encoding.
3399 *) Change the 'req' utility to use the new field handling and multibyte
3400 copy routines. Before the DN field creation was handled in an ad hoc
3401 way in req, ca, and x509 which was rather broken and didn't support
3402 BMPStrings or UTF8Strings. Since some software doesn't implement
3403 BMPStrings or UTF8Strings yet, they can be enabled using the config file
3404 using the dirstring_type option. See the new comment in the default
3405 openssl.cnf for more info.
3408 *) Make crypto/rand/md_rand.c more robust:
3409 - Assure unique random numbers after fork().
3410 - Make sure that concurrent threads access the global counter and
3411 md serializably so that we never lose entropy in them
3412 or use exactly the same state in multiple threads.
3413 Access to the large state is not always serializable because
3414 the additional locking could be a performance killer, and
3415 md should be large enough anyway.
3418 *) New file apps/app_rand.c with commonly needed functionality
3419 for handling the random seed file.
3421 Use the random seed file in some applications that previously did not:
3423 dsaparam -genkey (which also ignored its '-rand' option),
3426 x509 (when signing).
3427 Except on systems with /dev/urandom, it is crucial to have a random
3428 seed file at least for key creation, DSA signing, and for DH exchanges;
3429 for RSA signatures we could do without one.
3431 gendh and gendsa (unlike genrsa) used to read only the first byte
3432 of each file listed in the '-rand' option. The function as previously
3433 found in genrsa is now in app_rand.c and is used by all programs
3434 that support '-rand'.
3437 *) In RAND_write_file, use mode 0600 for creating files;
3438 don't just chmod when it may be too late.
3441 *) Report an error from X509_STORE_load_locations
3442 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
3445 *) New function ASN1_mbstring_copy() this copies a string in either
3446 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
3447 into an ASN1_STRING type. A mask of permissible types is passed
3448 and it chooses the "minimal" type to use or an error if not type
3452 *) Add function equivalents to the various macros in asn1.h. The old
3453 macros are retained with an M_ prefix. Code inside the library can
3454 use the M_ macros. External code (including the openssl utility)
3455 should *NOT* in order to be "shared library friendly".
3458 *) Add various functions that can check a certificate's extensions
3459 to see if it usable for various purposes such as SSL client,
3460 server or S/MIME and CAs of these types. This is currently
3461 VERY EXPERIMENTAL but will ultimately be used for certificate chain
3462 verification. Also added a -purpose flag to x509 utility to
3463 print out all the purposes.
3466 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
3470 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
3471 for, obtain and decode and extension and obtain its critical flag.
3472 This allows all the necessary extension code to be handled in a
3473 single function call.
3476 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
3477 platforms. See crypto/rc4/rc4_enc.c for further details.
3480 *) New -noout option to asn1parse. This causes no output to be produced
3481 its main use is when combined with -strparse and -out to extract data
3482 from a file (which may not be in ASN.1 format).
3485 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
3486 when producing the local key id.
3487 [Richard Levitte <levitte@stacken.kth.se>]
3489 *) New option -dhparam in s_server. This allows a DH parameter file to be
3490 stated explicitly. If it is not stated then it tries the first server
3491 certificate file. The previous behaviour hard coded the filename
3495 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
3496 a public key to be input or output. For example:
3497 openssl rsa -in key.pem -pubout -out pubkey.pem
3498 Also added necessary DSA public key functions to handle this.
3501 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
3502 in the message. This was handled by allowing
3503 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
3504 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
3506 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
3507 to the end of the strings whereas this didn't. This would cause problems
3508 if strings read with d2i_ASN1_bytes() were later modified.
3509 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
3511 *) Fix for base64 decode bug. When a base64 bio reads only one line of
3512 data and it contains EOF it will end up returning an error. This is
3513 caused by input 46 bytes long. The cause is due to the way base64
3514 BIOs find the start of base64 encoded data. They do this by trying a
3515 trial decode on each line until they find one that works. When they
3516 do a flag is set and it starts again knowing it can pass all the
3517 data directly through the decoder. Unfortunately it doesn't reset
3518 the context it uses. This means that if EOF is reached an attempt
3519 is made to pass two EOFs through the context and this causes the
3520 resulting error. This can also cause other problems as well. As is
3521 usual with these problems it takes *ages* to find and the fix is
3522 trivial: move one line.
3523 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
3525 *) Ugly workaround to get s_client and s_server working under Windows. The
3526 old code wouldn't work because it needed to select() on sockets and the
3527 tty (for keypresses and to see if data could be written). Win32 only
3528 supports select() on sockets so we select() with a 1s timeout on the
3529 sockets and then see if any characters are waiting to be read, if none
3530 are present then we retry, we also assume we can always write data to
3531 the tty. This isn't nice because the code then blocks until we've
3532 received a complete line of data and it is effectively polling the
3533 keyboard at 1s intervals: however it's quite a bit better than not
3534 working at all :-) A dedicated Windows application might handle this
3535 with an event loop for example.
3538 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
3539 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
3540 will be called when RSA_sign() and RSA_verify() are used. This is useful
3541 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
3542 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
3543 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
3544 This necessitated the support of an extra signature type NID_md5_sha1
3545 for SSL signatures and modifications to the SSL library to use it instead
3546 of calling RSA_public_decrypt() and RSA_private_encrypt().
3549 *) Add new -verify -CAfile and -CApath options to the crl program, these
3550 will lookup a CRL issuers certificate and verify the signature in a
3551 similar way to the verify program. Tidy up the crl program so it
3552 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
3553 less strict. It will now permit CRL extensions even if it is not
3554 a V2 CRL: this will allow it to tolerate some broken CRLs.
3557 *) Initialize all non-automatic variables each time one of the openssl
3558 sub-programs is started (this is necessary as they may be started
3559 multiple times from the "OpenSSL>" prompt).
3560 [Lennart Bang, Bodo Moeller]
3562 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
3563 removing all other RSA functionality (this is what NO_RSA does). This
3564 is so (for example) those in the US can disable those operations covered
3565 by the RSA patent while allowing storage and parsing of RSA keys and RSA
3569 *) Non-copying interface to BIO pairs.
3570 (still largely untested)
3573 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
3574 ASCII string. This was handled independently in various places before.
3577 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
3578 UTF8 strings a character at a time.
3581 *) Use client_version from client hello to select the protocol
3582 (s23_srvr.c) and for RSA client key exchange verification
3583 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
3586 *) Add various utility functions to handle SPKACs, these were previously
3587 handled by poking round in the structure internals. Added new function
3588 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
3589 print, verify and generate SPKACs. Based on an original idea from
3590 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
3593 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
3596 *) Allow the config file extension section to be overwritten on the
3597 command line. Based on an original idea from Massimiliano Pala
3598 <madwolf@comune.modena.it>. The new option is called -extensions
3599 and can be applied to ca, req and x509. Also -reqexts to override
3600 the request extensions in req and -crlexts to override the crl extensions
3604 *) Add new feature to the SPKAC handling in ca. Now you can include
3605 the same field multiple times by preceding it by "XXXX." for example:
3608 this is the same syntax as used in the req config file.
3611 *) Allow certificate extensions to be added to certificate requests. These
3612 are specified in a 'req_extensions' option of the req section of the
3613 config file. They can be printed out with the -text option to req but
3614 are otherwise ignored at present.
3617 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
3618 data read consists of only the final block it would not decrypted because
3619 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
3620 A misplaced 'break' also meant the decrypted final block might not be
3621 copied until the next read.
3624 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
3625 a few extra parameters to the DH structure: these will be useful if
3626 for example we want the value of 'q' or implement X9.42 DH.
3629 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
3630 provides hooks that allow the default DSA functions or functions on a
3631 "per key" basis to be replaced. This allows hardware acceleration and
3632 hardware key storage to be handled without major modification to the
3633 library. Also added low level modexp hooks and CRYPTO_EX structure and
3634 associated functions.
3637 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
3638 as "read only": it can't be written to and the buffer it points to will
3639 not be freed. Reading from a read only BIO is much more efficient than
3640 a normal memory BIO. This was added because there are several times when
3641 an area of memory needs to be read from a BIO. The previous method was
3642 to create a memory BIO and write the data to it, this results in two
3643 copies of the data and an O(n^2) reading algorithm. There is a new
3644 function BIO_new_mem_buf() which creates a read only memory BIO from
3645 an area of memory. Also modified the PKCS#7 routines to use read only
3649 *) Bugfix: ssl23_get_client_hello did not work properly when called in
3650 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
3651 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
3652 but a retry condition occured while trying to read the rest.
3655 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
3656 NID_pkcs7_encrypted by default: this was wrong since this should almost
3657 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
3658 the encrypted data type: this is a more sensible place to put it and it
3659 allows the PKCS#12 code to be tidied up that duplicated this
3663 *) Changed obj_dat.pl script so it takes its input and output files on
3664 the command line. This should avoid shell escape redirection problems
3668 *) Initial support for certificate extension requests, these are included
3669 in things like Xenroll certificate requests. Included functions to allow
3670 extensions to be obtained and added.
3673 *) -crlf option to s_client and s_server for sending newlines as
3674 CRLF (as required by many protocols).
3677 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
3679 *) Install libRSAglue.a when OpenSSL is built with RSAref.
3680 [Ralf S. Engelschall]
3682 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
3683 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
3685 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
3689 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
3690 DH parameters/keys (q is lost during that conversion, but the resulting
3691 DH parameters contain its length).
3693 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
3694 much faster than DH_generate_parameters (which creates parameters
3695 where p = 2*q + 1), and also the smaller q makes DH computations
3696 much more efficient (160-bit exponentiation instead of 1024-bit
3697 exponentiation); so this provides a convenient way to support DHE
3698 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
3699 utter importance to use
3700 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
3702 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
3703 when such DH parameters are used, because otherwise small subgroup
3704 attacks may become possible!
3707 *) Avoid memory leak in i2d_DHparams.
3710 *) Allow the -k option to be used more than once in the enc program:
3711 this allows the same encrypted message to be read by multiple recipients.
3714 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
3715 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
3716 it will always use the numerical form of the OID, even if it has a short
3720 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
3721 method only got called if p,q,dmp1,dmq1,iqmp components were present,
3722 otherwise bn_mod_exp was called. In the case of hardware keys for example
3723 no private key components need be present and it might store extra data
3724 in the RSA structure, which cannot be accessed from bn_mod_exp.
3725 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
3726 private key operations.
3729 *) Added support for SPARC Linux.
3732 *) pem_password_cb function type incompatibly changed from
3733 typedef int pem_password_cb(char *buf, int size, int rwflag);
3735 ....(char *buf, int size, int rwflag, void *userdata);
3736 so that applications can pass data to their callbacks:
3737 The PEM[_ASN1]_{read,write}... functions and macros now take an
3738 additional void * argument, which is just handed through whenever
3739 the password callback is called.
3740 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
3742 New function SSL_CTX_set_default_passwd_cb_userdata.
3744 Compatibility note: As many C implementations push function arguments
3745 onto the stack in reverse order, the new library version is likely to
3746 interoperate with programs that have been compiled with the old
3747 pem_password_cb definition (PEM_whatever takes some data that
3748 happens to be on the stack as its last argument, and the callback
3749 just ignores this garbage); but there is no guarantee whatsoever that
3752 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
3753 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
3754 problems not only on Windows, but also on some Unix platforms.
3755 To avoid problematic command lines, these definitions are now in an
3756 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
3757 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
3760 *) MIPS III/IV assembler module is reimplemented.
3763 *) More DES library cleanups: remove references to srand/rand and
3764 delete an unused file.
3767 *) Add support for the the free Netwide assembler (NASM) under Win32,
3768 since not many people have MASM (ml) and it can be hard to obtain.
3769 This is currently experimental but it seems to work OK and pass all
3770 the tests. Check out INSTALL.W32 for info.
3773 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
3774 without temporary keys kept an extra copy of the server key,
3775 and connections with temporary keys did not free everything in case
3779 *) New function RSA_check_key and new openssl rsa option -check
3780 for verifying the consistency of RSA keys.
3781 [Ulf Moeller, Bodo Moeller]
3783 *) Various changes to make Win32 compile work:
3784 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
3785 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
3786 comparison" warnings.
3787 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
3790 *) Add a debugging option to PKCS#5 v2 key generation function: when
3791 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
3792 derived keys are printed to stderr.
3795 *) Copy the flags in ASN1_STRING_dup().
3796 [Roman E. Pavlov <pre@mo.msk.ru>]
3798 *) The x509 application mishandled signing requests containing DSA
3799 keys when the signing key was also DSA and the parameters didn't match.
3801 It was supposed to omit the parameters when they matched the signing key:
3802 the verifying software was then supposed to automatically use the CA's
3803 parameters if they were absent from the end user certificate.
3805 Omitting parameters is no longer recommended. The test was also
3806 the wrong way round! This was probably due to unusual behaviour in
3807 EVP_cmp_parameters() which returns 1 if the parameters match.
3808 This meant that parameters were omitted when they *didn't* match and
3809 the certificate was useless. Certificates signed with 'ca' didn't have
3811 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
3813 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
3814 The interface is as follows:
3815 Applications can use
3816 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
3817 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
3818 "off" is now the default.
3819 The library internally uses
3820 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
3821 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
3822 to disable memory-checking temporarily.
3824 Some inconsistent states that previously were possible (and were
3825 even the default) are now avoided.
3827 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
3828 with each memory chunk allocated; this is occasionally more helpful
3829 than just having a counter.
3831 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
3833 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
3837 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
3838 which largely parallels "options", but is for changing API behaviour,
3839 whereas "options" are about protocol behaviour.
3840 Initial "mode" flags are:
3842 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
3843 a single record has been written.
3844 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
3845 retries use the same buffer location.
3846 (But all of the contents must be
3850 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
3853 *) Fix problems with no-hmac etc.
3854 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
3856 *) New functions RSA_get_default_method(), RSA_set_method() and
3857 RSA_get_method(). These allows replacement of RSA_METHODs without having
3858 to mess around with the internals of an RSA structure.
3861 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
3862 Also really enable memory leak checks in openssl.c and in some
3864 [Chad C. Mulligan, Bodo Moeller]
3866 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
3867 up the length of negative integers. This has now been simplified to just
3868 store the length when it is first determined and use it later, rather
3869 than trying to keep track of where data is copied and updating it to
3871 [Steve Henson, reported by Brien Wheeler
3872 <bwheeler@authentica-security.com>]
3874 *) Add a new function PKCS7_signatureVerify. This allows the verification
3875 of a PKCS#7 signature but with the signing certificate passed to the
3876 function itself. This contrasts with PKCS7_dataVerify which assumes the
3877 certificate is present in the PKCS#7 structure. This isn't always the
3878 case: certificates can be omitted from a PKCS#7 structure and be
3879 distributed by "out of band" means (such as a certificate database).
3882 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
3883 function prototypes in pem.h, also change util/mkdef.pl to add the
3884 necessary function names.
3887 *) mk1mf.pl (used by Windows builds) did not properly read the
3888 options set by Configure in the top level Makefile, and Configure
3889 was not even able to write more than one option correctly.
3890 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
3893 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
3894 file to be loaded from a BIO or FILE pointer. The BIO version will
3895 for example allow memory BIOs to contain config info.
3898 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
3899 Whoever hopes to achieve shared-library compatibility across versions
3900 must use this, not the compile-time macro.
3901 (Exercise 0.9.4: Which is the minimum library version required by
3903 Note: All this applies only to multi-threaded programs, others don't
3907 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
3908 through a BIO pair triggered the default case, i.e.
3909 SSLerr(...,SSL_R_UNKNOWN_STATE).
3912 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
3913 can use the SSL library even if none of the specific BIOs is
3917 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
3918 for the encoded length.
3919 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
3921 *) Add initial documentation of the X509V3 functions.
3924 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
3925 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
3926 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
3927 secure PKCS#8 private key format with a high iteration count.
3930 *) Fix determination of Perl interpreter: A perl or perl5
3931 _directory_ in $PATH was also accepted as the interpreter.
3932 [Ralf S. Engelschall]
3934 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
3935 wrong with it but it was very old and did things like calling
3936 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
3940 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
3941 to use the new extension code.
3944 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
3945 with macros. This should make it easier to change their form, add extra
3946 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
3950 *) Add to configuration table a new entry that can specify an alternative
3951 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
3952 according to Mark Crispin <MRC@Panda.COM>.
3956 *) DES CBC did not update the IV. Weird.
3959 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
3960 Changing the behaviour of the former might break existing programs --
3961 where IV updating is needed, des_ncbc_encrypt can be used.
3964 *) When bntest is run from "make test" it drives bc to check its
3965 calculations, as well as internally checking them. If an internal check
3966 fails, it needs to cause bc to give a non-zero result or make test carries
3967 on without noticing the failure. Fixed.
3970 *) DES library cleanups.
3973 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
3974 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
3975 ciphers. NOTE: although the key derivation function has been verified
3976 against some published test vectors it has not been extensively tested
3977 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
3981 *) Instead of "mkdir -p", which is not fully portable, use new
3982 Perl script "util/mkdir-p.pl".
3985 *) Rewrite the way password based encryption (PBE) is handled. It used to
3986 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
3987 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
3988 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
3989 the 'parameter' field of the AlgorithmIdentifier is passed to the
3990 underlying key generation function so it must do its own ASN1 parsing.
3991 This has also changed the EVP_PBE_CipherInit() function which now has a
3992 'parameter' argument instead of literal salt and iteration count values
3993 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
3996 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
3997 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
3998 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
3999 KEY" because this clashed with PKCS#8 unencrypted string. Since this
4000 value was just used as a "magic string" and not used directly its
4001 value doesn't matter.
4004 *) Introduce some semblance of const correctness to BN. Shame C doesn't
4008 *) "linux-sparc64" configuration (ultrapenguin).
4009 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
4010 "linux-sparc" configuration.
4011 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
4013 *) config now generates no-xxx options for missing ciphers.
4016 *) Support the EBCDIC character set (work in progress).
4017 File ebcdic.c not yet included because it has a different license.
4018 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
4020 *) Support BS2000/OSD-POSIX.
4021 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
4023 *) Make callbacks for key generation use void * instead of char *.
4026 *) Make S/MIME samples compile (not yet tested).
4029 *) Additional typesafe stacks.
4032 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
4036 Changes between 0.9.3 and 0.9.3a [29 May 1999]
4038 *) New configuration variant "sco5-gcc".
4040 *) Updated some demos.
4041 [Sean O Riordain, Wade Scholine]
4043 *) Add missing BIO_free at exit of pkcs12 application.
4046 *) Fix memory leak in conf.c.
4049 *) Updates for Win32 to assembler version of MD5.
4052 *) Set #! path to perl in apps/der_chop to where we found it
4053 instead of using a fixed path.
4056 *) SHA library changes for irix64-mips4-cc.
4059 *) Improvements for VMS support.
4063 Changes between 0.9.2b and 0.9.3 [24 May 1999]
4065 *) Bignum library bug fix. IRIX 6 passes "make test" now!
4066 This also avoids the problems with SC4.2 and unpatched SC5.
4067 [Andy Polyakov <appro@fy.chalmers.se>]
4069 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
4070 These are required because of the typesafe stack would otherwise break
4071 existing code. If old code used a structure member which used to be STACK
4072 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
4073 sk_num or sk_value it would produce an error because the num, data members
4074 are not present in STACK_OF. Now it just produces a warning. sk_set
4075 replaces the old method of assigning a value to sk_value
4076 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
4077 that does this will no longer work (and should use sk_set instead) but
4078 this could be regarded as a "questionable" behaviour anyway.
4081 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
4082 correctly handle encrypted S/MIME data.
4085 *) Change type of various DES function arguments from des_cblock
4086 (which means, in function argument declarations, pointer to char)
4087 to des_cblock * (meaning pointer to array with 8 char elements),
4088 which allows the compiler to do more typechecking; it was like
4089 that back in SSLeay, but with lots of ugly casts.
4091 Introduce new type const_des_cblock.
4094 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
4095 problems: find RecipientInfo structure that matches recipient certificate
4096 and initialise the ASN1 structures properly based on passed cipher.
4099 *) Belatedly make the BN tests actually check the results.
4102 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
4103 to and from BNs: it was completely broken. New compilation option
4104 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
4105 key elements as negative integers.
4108 *) Reorganize and speed up MD5.
4109 [Andy Polyakov <appro@fy.chalmers.se>]
4112 [Richard Levitte <richard@levitte.org>]
4114 *) New option -out to asn1parse to allow the parsed structure to be
4115 output to a file. This is most useful when combined with the -strparse
4116 option to examine the output of things like OCTET STRINGS.
4119 *) Make SSL library a little more fool-proof by not requiring any longer
4120 that SSL_set_{accept,connect}_state be called before
4121 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
4122 in many applications because usually everything *appeared* to work as
4123 intended anyway -- now it really works as intended).
4126 *) Move openssl.cnf out of lib/.
4129 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
4130 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
4131 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
4132 [Ralf S. Engelschall]
4134 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
4135 handle PKCS#7 enveloped data properly.
4136 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
4138 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
4139 copying pointers. The cert_st handling is changed by this in
4140 various ways (and thus what used to be known as ctx->default_cert
4141 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
4142 any longer when s->cert does not give us what we need).
4143 ssl_cert_instantiate becomes obsolete by this change.
4144 As soon as we've got the new code right (possibly it already is?),
4145 we have solved a couple of bugs of the earlier code where s->cert
4146 was used as if it could not have been shared with other SSL structures.
4148 Note that using the SSL API in certain dirty ways now will result
4149 in different behaviour than observed with earlier library versions:
4150 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
4151 does not influence s as it used to.
4153 In order to clean up things more thoroughly, inside SSL_SESSION
4154 we don't use CERT any longer, but a new structure SESS_CERT
4155 that holds per-session data (if available); currently, this is
4156 the peer's certificate chain and, for clients, the server's certificate
4157 and temporary key. CERT holds only those values that can have
4158 meaningful defaults in an SSL_CTX.
4161 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
4162 from the internal representation. Various PKCS#7 fixes: remove some
4163 evil casts and set the enc_dig_alg field properly based on the signing
4167 *) Allow PKCS#12 password to be set from the command line or the
4168 environment. Let 'ca' get its config file name from the environment
4169 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
4173 *) Allow certificate policies extension to use an IA5STRING for the
4174 organization field. This is contrary to the PKIX definition but
4175 VeriSign uses it and IE5 only recognises this form. Document 'x509'
4179 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
4180 without disallowing inline assembler and the like for non-pedantic builds.
4183 *) Support Borland C++ builder.
4184 [Janez Jere <jj@void.si>, modified by Ulf Möller]
4189 *) SHA-1 cleanups and performance enhancements.
4190 [Andy Polyakov <appro@fy.chalmers.se>]
4192 *) Sparc v8plus assembler for the bignum library.
4193 [Andy Polyakov <appro@fy.chalmers.se>]
4195 *) Accept any -xxx and +xxx compiler options in Configure.
4198 *) Update HPUX configuration.
4201 *) Add missing sk_<type>_unshift() function to safestack.h
4202 [Ralf S. Engelschall]
4204 *) New function SSL_CTX_use_certificate_chain_file that sets the
4205 "extra_cert"s in addition to the certificate. (This makes sense
4206 only for "PEM" format files, as chains as a whole are not
4210 *) Support verify_depth from the SSL API.
4211 x509_vfy.c had what can be considered an off-by-one-error:
4212 Its depth (which was not part of the external interface)
4213 was actually counting the number of certificates in a chain;
4214 now it really counts the depth.
4217 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
4218 instead of X509err, which often resulted in confusing error
4219 messages since the error codes are not globally unique
4220 (e.g. an alleged error in ssl3_accept when a certificate
4221 didn't match the private key).
4223 *) New function SSL_CTX_set_session_id_context that allows to set a default
4224 value (so that you don't need SSL_set_session_id_context for each
4225 connection using the SSL_CTX).
4228 *) OAEP decoding bug fix.
4231 *) Support INSTALL_PREFIX for package builders, as proposed by
4235 *) New Configure options "threads" and "no-threads". For systems
4236 where the proper compiler options are known (currently Solaris
4237 and Linux), "threads" is the default.
4240 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
4243 *) Install various scripts to $(OPENSSLDIR)/misc, not to
4244 $(INSTALLTOP)/bin -- they shouldn't clutter directories
4245 such as /usr/local/bin.
4248 *) "make linux-shared" to build shared libraries.
4249 [Niels Poppe <niels@netbox.org>]
4251 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
4254 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
4255 extension adding in x509 utility.
4258 *) Remove NOPROTO sections and error code comments.
4261 *) Partial rewrite of the DEF file generator to now parse the ANSI
4265 *) New Configure options --prefix=DIR and --openssldir=DIR.
4268 *) Complete rewrite of the error code script(s). It is all now handled
4269 by one script at the top level which handles error code gathering,
4270 header rewriting and C source file generation. It should be much better
4271 than the old method: it now uses a modified version of Ulf's parser to
4272 read the ANSI prototypes in all header files (thus the old K&R definitions
4273 aren't needed for error creation any more) and do a better job of
4274 translating function codes into names. The old 'ASN1 error code imbedded
4275 in a comment' is no longer necessary and it doesn't use .err files which
4276 have now been deleted. Also the error code call doesn't have to appear all
4277 on one line (which resulted in some large lines...).
4280 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
4283 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
4284 0 (which usually indicates a closed connection), but continue reading.
4287 *) Fix some race conditions.
4290 *) Add support for CRL distribution points extension. Add Certificate
4291 Policies and CRL distribution points documentation.
4294 *) Move the autogenerated header file parts to crypto/opensslconf.h.
4297 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
4298 8 of keying material. Merlin has also confirmed interop with this fix
4299 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
4300 [Merlin Hughes <merlin@baltimore.ie>]
4302 *) Fix lots of warnings.
4303 [Richard Levitte <levitte@stacken.kth.se>]
4305 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
4306 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
4307 [Richard Levitte <levitte@stacken.kth.se>]
4309 *) Fix problems with sizeof(long) == 8.
4310 [Andy Polyakov <appro@fy.chalmers.se>]
4312 *) Change functions to ANSI C.
4315 *) Fix typos in error codes.
4316 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
4318 *) Remove defunct assembler files from Configure.
4321 *) SPARC v8 assembler BIGNUM implementation.
4322 [Andy Polyakov <appro@fy.chalmers.se>]
4324 *) Support for Certificate Policies extension: both print and set.
4325 Various additions to support the r2i method this uses.
4328 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
4329 return a const string when you are expecting an allocated buffer.
4332 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
projects / openssl.git / blob