5 Changes between 0.9.5a and 0.9.6 [xx XXX 2000]
7 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
8 This makes DSA verification about 2 % faster.
11 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
12 (meaning that now 2^5 values will be precomputed, which is only 4 KB
13 plus overhead for 1024 bit moduli).
14 This makes exponentiations about 0.5 % faster for 1024 bit
15 exponents (as measured by "openssl speed rsa2048").
18 *) Rename memory handling macros to avoid conflicts with other
20 Malloc => OPENSSL_malloc
21 Malloc_locked => OPENSSL_malloc_locked
22 Realloc => OPENSSL_realloc
26 *) New function BN_mod_exp_mont_word for small bases (roughly 15-20%
27 faster than BN_mod_exp_mont, i.e. 7-10% for a full DH exchange).
31 [John Jarvie <jjarvie@newsguy.com>]
33 *) The type-safe stack code has been rejigged. It is now only compiled
34 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
35 by default all type-specific stack functions are "#define"d back to
36 standard stack functions. This results in more streamlined output
37 but retains the type-safety checking possibilities of the original
41 *) The STACK code has been cleaned up, and certain type declarations
42 that didn't make a lot of sense have been brought in line. This has
43 also involved a cleanup of sorts in safestack.h to more correctly
44 map type-safe stack functions onto their plain stack counterparts.
45 This work has also resulted in a variety of "const"ifications of
46 lots of the code, especially "_cmp" operations which should normally
47 be prototyped with "const" parameters anyway.
50 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
51 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
52 (The PRNG state consists of two parts, the large pool 'state' and 'md',
53 where all of 'md' is used each time the PRNG is used, but 'state'
54 is used only indexed by a cyclic counter. As entropy may not be
55 well distributed from the beginning, 'md' is important as a
56 chaining variable. However, the output function chains only half
57 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
58 all of 'md', and seeding with STATE_SIZE dummy bytes will result
59 in all of 'state' being rewritten, with the new values depending
60 on virtually all of 'md'. This overcomes the 80 bit limitation.)
63 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
64 the handshake is continued after ssl_verify_cert_chain();
65 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
66 can lead to 'unexplainable' connection aborts later.
67 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
69 *) Major EVP API cipher revision.
70 Add hooks for extra EVP features. This allows various cipher
71 parameters to be set in the EVP interface. Support added for variable
72 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
73 setting of RC2 and RC5 parameters.
75 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
78 Remove lots of duplicated code from the EVP library. For example *every*
79 cipher init() function handles the 'iv' in the same way according to the
80 cipher mode. They also all do nothing if the 'key' parameter is NULL and
81 for CFB and OFB modes they zero ctx->num.
83 New functionality allows removal of S/MIME code RC2 hack.
85 Most of the routines have the same form and so can be declared in terms
88 By shifting this to the top level EVP_CipherInit() it can be removed from
89 all individual ciphers. If the cipher wants to handle IVs or keys
90 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
93 Change lots of functions like EVP_EncryptUpdate() to now return a
94 value: although software versions of the algorithms cannot fail
95 any installed hardware versions can.
98 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
99 this option is set, tolerate broken clients that send the negotiated
100 protocol version number instead of the requested protocol version
104 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
105 i.e. non-zero for export ciphersuites, zero otherwise.
106 Previous versions had this flag inverted, inconsistent with
107 rsa_tmp_cb (..._TMP_RSA_CB).
108 [Bodo Moeller; problem reported by Amit Chopra]
110 *) Add missing DSA library text string. Work around for some IIS
111 key files with invalid SEQUENCE encoding.
114 *) Add a document (doc/standards.txt) that list all kinds of standards
115 and so on that are implemented in OpenSSL.
118 *) Enhance c_rehash script. Old version would mishandle certificates
119 with the same subject name hash and wouldn't handle CRLs at all.
120 Added -fingerprint option to crl utility, to support new c_rehash
124 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
127 *) Fix for SSL server purpose checking. Server checking was
128 rejecting certificates which had extended key usage present
129 but no ssl client purpose.
130 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
132 *) Make PKCS#12 code work with no password. The PKCS#12 spec
133 is a little unclear about how a blank password is handled.
134 Since the password in encoded as a BMPString with terminating
135 double NULL a zero length password would end up as just the
136 double NULL. However no password at all is different and is
137 handled differently in the PKCS#12 key generation code. NS
138 treats a blank password as zero length. MSIE treats it as no
139 password on export: but it will try both on import. We now do
140 the same: PKCS12_parse() tries zero length and no password if
141 the password is set to "" or NULL (NULL is now a valid password:
142 it wasn't before) as does the pkcs12 application.
145 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
146 perror when PEM_read_bio_X509_REQ fails, the error message must
147 be obtained from the error queue.
150 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
151 it in ERR_remove_state if appropriate, and change ERR_get_state
152 accordingly to avoid race conditions (this is necessary because
153 thread_hash is no longer constant once set).
156 *) Bugfix for linux-elf makefile.one.
159 *) RSA_get_default_method() will now cause a default
160 RSA_METHOD to be chosen if one doesn't exist already.
161 Previously this was only set during a call to RSA_new()
162 or RSA_new_method(NULL) meaning it was possible for
163 RSA_get_default_method() to return NULL.
166 *) Added native name translation to the existing DSO code
167 that will convert (if the flag to do so is set) filenames
168 that are sufficiently small and have no path information
169 into a canonical native form. Eg. "blah" converted to
170 "libblah.so" or "blah.dll" etc.
173 *) New function ERR_error_string_n(e, buf, len) which is like
174 ERR_error_string(e, buf), but writes at most 'len' bytes
175 including the 0 terminator. For ERR_error_string_n, 'buf'
177 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
179 *) CONF library reworked to become more general. A new CONF
180 configuration file reader "class" is implemented as well as a
181 new functions (NCONF_*, for "New CONF") to handle it. The now
182 old CONF_* functions are still there, but are reimplemented to
183 work in terms of the new functions. Also, a set of functions
184 to handle the internal storage of the configuration data is
185 provided to make it easier to write new configuration file
186 reader "classes" (I can definitely see something reading a
187 configuration file in XML format, for example), called _CONF_*,
188 or "the configuration storage API"...
190 The new configuration file reading functions are:
192 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
193 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
195 NCONF_default, NCONF_WIN32
197 NCONF_dump_fp, NCONF_dump_bio
199 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
200 NCONF_new creates a new CONF object. This works in the same way
201 as other interfaces in OpenSSL, like the BIO interface.
202 NCONF_dump_* dump the internal storage of the configuration file,
203 which is useful for debugging. All other functions take the same
204 arguments as the old CONF_* functions wth the exception of the
205 first that must be a `CONF *' instead of a `LHASH *'.
207 To make it easer to use the new classes with the old CONF_* functions,
208 the function CONF_set_default_method is provided.
211 *) Add '-tls1' option to 'openssl ciphers', which was already
212 mentioned in the documentation but had not been implemented.
213 (This option is not yet really useful because even the additional
214 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
217 *) Initial DSO code added into libcrypto for letting OpenSSL (and
218 OpenSSL-based applications) load shared libraries and bind to
219 them in a portable way.
220 [Geoff Thorpe, with contributions from Richard Levitte]
222 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
224 *) Make sure _lrotl and _lrotr are only used with MSVC.
226 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
227 (the default implementation of RAND_status).
229 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
230 to '-clrext' (= clear extensions), as intended and documented.
231 [Bodo Moeller; inconsistency pointed out by Michael Attili
234 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
235 was larger than the MD block size.
236 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
238 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
239 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
240 using the passed key: if the passed key was a private key the result
241 of X509_print(), for example, would be to print out all the private key
245 *) des_quad_cksum() byte order bug fix.
246 [Ulf Möller, using the problem description in krb4-0.9.7, where
247 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
249 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
251 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
253 *) For easily testing in shell scripts whether some command
254 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
255 returns with exit code 0 iff no command of the given name is available.
256 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
257 the output goes to stdout and nothing is printed to stderr.
258 Additional arguments are always ignored.
260 Since for each cipher there is a command of the same name,
261 the 'no-cipher' compilation switches can be tested this way.
263 ('openssl no-XXX' is not able to detect pseudo-commands such
264 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
267 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
270 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
271 is set; it will be thrown away anyway because each handshake creates
273 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
274 to parameters -- in previous versions (since OpenSSL 0.9.3) the
275 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
276 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
279 *) New s_client option -ign_eof: EOF at stdin is ignored, and
280 'Q' and 'R' lose their special meanings (quit/renegotiate).
281 This is part of what -quiet does; unlike -quiet, -ign_eof
282 does not suppress any output.
285 *) Add compatibility options to the purpose and trust code. The
286 purpose X509_PURPOSE_ANY is "any purpose" which automatically
287 accepts a certificate or CA, this was the previous behaviour,
288 with all the associated security issues.
290 X509_TRUST_COMPAT is the old trust behaviour: only and
291 automatically trust self signed roots in certificate store. A
292 new trust setting X509_TRUST_DEFAULT is used to specify that
293 a purpose has no associated trust setting and it should instead
294 use the value in the default purpose.
297 *) Fix the PKCS#8 DSA private key code so it decodes keys again
298 and fix a memory leak.
301 *) In util/mkerr.pl (which implements 'make errors'), preserve
302 reason strings from the previous version of the .c file, as
303 the default to have only downcase letters (and digits) in
304 automatically generated reasons codes is not always appropriate.
307 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
308 using strerror. Previously, ERR_reason_error_string() returned
309 library names as reason strings for SYSerr; but SYSerr is a special
310 case where small numbers are errno values, not library numbers.
313 *) Add '-dsaparam' option to 'openssl dhparam' application. This
314 converts DSA parameters into DH parameters. (When creating parameters,
315 DSA_generate_parameters is used.)
318 *) Include 'length' (recommended exponent length) in C code generated
319 by 'openssl dhparam -C'.
322 *) The second argument to set_label in perlasm was already being used
323 so couldn't be used as a "file scope" flag. Moved to third argument
327 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
328 instead of RAND_bytes for encryption IVs and salts.
331 *) Include RAND_status() into RAND_METHOD instead of implementing
332 it only for md_rand.c Otherwise replacing the PRNG by calling
333 RAND_set_rand_method would be impossible.
336 *) Don't let DSA_generate_key() enter an infinite loop if the random
337 number generation fails.
340 *) New 'rand' application for creating pseudo-random output.
343 *) Added configuration support for Linux/IA64
344 [Rolf Haberrecker <rolf@suse.de>]
346 *) Assembler module support for Mingw32.
349 *) Shared library support for HPUX (in shlib/).
350 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
352 *) Shared library support for Solaris gcc.
353 [Lutz Behnke <behnke@trustcenter.de>]
355 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
357 *) PKCS7_encrypt() was adding text MIME headers twice because they
358 were added manually and by SMIME_crlf_copy().
361 *) In bntest.c don't call BN_rand with zero bits argument.
362 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
364 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
365 case was implemented. This caused BN_div_recp() to fail occasionally.
368 *) Add an optional second argument to the set_label() in the perl
369 assembly language builder. If this argument exists and is set
370 to 1 it signals that the assembler should use a symbol whose
371 scope is the entire file, not just the current function. This
372 is needed with MASM which uses the format label:: for this scope.
373 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
375 *) Change the ASN1 types so they are typedefs by default. Before
376 almost all types were #define'd to ASN1_STRING which was causing
377 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
381 *) Change names of new functions to the new get1/get0 naming
382 convention: After 'get1', the caller owns a reference count
383 and has to call ..._free; 'get0' returns a pointer to some
384 data structure without incrementing reference counters.
385 (Some of the existing 'get' functions increment a reference
386 counter, some don't.)
387 Similarly, 'set1' and 'add1' functions increase reference
388 counters or duplicate objects.
391 *) Allow for the possibility of temp RSA key generation failure:
392 the code used to assume it always worked and crashed on failure.
395 *) Fix potential buffer overrun problem in BIO_printf().
396 [Ulf Möller, using public domain code by Patrick Powell; problem
397 pointed out by David Sacerdote <das33@cornell.edu>]
399 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
400 RAND_egd() and RAND_status(). In the command line application,
401 the EGD socket can be specified like a seed file using RANDFILE
405 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
406 Some CAs (e.g. Verisign) distribute certificates in this form.
409 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
410 list to exclude them. This means that no special compilation option
411 is needed to use anonymous DH: it just needs to be included in the
415 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
416 EVP_MD_type. The old functionality is available in a new macro called
417 EVP_MD_md(). Change code that uses it and update docs.
420 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
421 where the 'void *' argument is replaced by a function pointer argument.
422 Previously 'void *' was abused to point to functions, which works on
423 many platforms, but is not correct. As these functions are usually
424 called by macros defined in OpenSSL header files, most source code
425 should work without changes.
428 *) <openssl/opensslconf.h> (which is created by Configure) now contains
429 sections with information on -D... compiler switches used for
430 compiling the library so that applications can see them. To enable
431 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
432 must be defined. E.g.,
433 #define OPENSSL_ALGORITHM_DEFINES
434 #include <openssl/opensslconf.h>
435 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
436 [Richard Levitte, Ulf and Bodo Möller]
438 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
442 *) Change the 'other' type in certificate aux info to a STACK_OF
443 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
444 the required ASN1 format: arbitrary types determined by an OID.
447 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
448 argument to 'req'. This is not because the function is newer or
449 better than others it just uses the work 'NEW' in the certificate
450 request header lines. Some software needs this.
453 *) Reorganise password command line arguments: now passwords can be
454 obtained from various sources. Delete the PEM_cb function and make
455 it the default behaviour: i.e. if the callback is NULL and the
456 usrdata argument is not NULL interpret it as a null terminated pass
457 phrase. If usrdata and the callback are NULL then the pass phrase
458 is prompted for as usual.
461 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
462 the support is automatically enabled. The resulting binaries will
463 autodetect the card and use it if present.
464 [Ben Laurie and Compaq Inc.]
466 *) Work around for Netscape hang bug. This sends certificate request
467 and server done in one record. Since this is perfectly legal in the
468 SSL/TLS protocol it isn't a "bug" option and is on by default. See
469 the bugs/SSLv3 entry for more info.
472 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
475 *) Add -rand argument to smime and pkcs12 applications and read/write
479 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
482 *) Add command line password options to the remaining applications.
485 *) Bug fix for BN_div_recp() for numerators with an even number of
489 *) More tests in bntest.c, and changed test_bn output.
492 *) ./config recognizes MacOS X now.
495 *) Bug fix for BN_div() when the first words of num and divsor are
496 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
499 *) Add support for various broken PKCS#8 formats, and command line
500 options to produce them.
503 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
504 get temporary BIGNUMs from a BN_CTX.
507 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
511 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
512 include a #define from the old name to the new. The original intent
513 was that statically linked binaries could for example just call
514 SSLeay_add_all_ciphers() to just add ciphers to the table and not
515 link with digests. This never worked becayse SSLeay_add_all_digests()
516 and SSLeay_add_all_ciphers() were in the same source file so calling
517 one would link with the other. They are now in separate source files.
520 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
523 *) Use a less unusual form of the Miller-Rabin primality test (it used
524 a binary algorithm for exponentiation integrated into the Miller-Rabin
525 loop, our standard modexp algorithms are faster).
528 *) Support for the EBCDIC character set completed.
529 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
531 *) Source code cleanups: use const where appropriate, eliminate casts,
532 use void * instead of char * in lhash.
535 *) Bugfix: ssl3_send_server_key_exchange was not restartable
536 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
537 this the server could overwrite ephemeral keys that the client
541 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
542 using 50 iterations of the Rabin-Miller test.
544 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
545 iterations of the Rabin-Miller test as required by the appendix
546 to FIPS PUB 186[-1]) instead of DSA_is_prime.
547 As BN_is_prime_fasttest includes trial division, DSA parameter
548 generation becomes much faster.
550 This implies a change for the callback functions in DSA_is_prime
551 and DSA_generate_parameters: The callback function is called once
552 for each positive witness in the Rabin-Miller test, not just
553 occasionally in the inner loop; and the parameters to the
554 callback function now provide an iteration count for the outer
555 loop rather than for the current invocation of the inner loop.
556 DSA_generate_parameters additionally can call the callback
557 function with an 'iteration count' of -1, meaning that a
558 candidate has passed the trial division test (when q is generated
559 from an application-provided seed, trial division is skipped).
562 *) New function BN_is_prime_fasttest that optionally does trial
563 division before starting the Rabin-Miller test and has
564 an additional BN_CTX * argument (whereas BN_is_prime always
565 has to allocate at least one BN_CTX).
566 'callback(1, -1, cb_arg)' is called when a number has passed the
567 trial division stage.
570 *) Fix for bug in CRL encoding. The validity dates weren't being handled
574 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
577 *) New function BN_pseudo_rand().
580 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
581 bignum version of BN_from_montgomery() with the working code from
582 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
586 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
587 made it impossible to use the same SSL_SESSION data structure in
588 SSL2 clients in multiple threads.
591 *) The return value of RAND_load_file() no longer counts bytes obtained
592 by stat(). RAND_load_file(..., -1) is new and uses the complete file
593 to seed the PRNG (previously an explicit byte count was required).
594 [Ulf Möller, Bodo Möller]
596 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
597 used (char *) instead of (void *) and had casts all over the place.
600 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
603 *) Retain source code compatibility for BN_prime_checks macro:
604 BN_is_prime(..., BN_prime_checks, ...) now uses
605 BN_prime_checks_for_size to determine the appropriate number of
606 Rabin-Miller iterations.
609 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
610 DH_CHECK_P_NOT_SAFE_PRIME.
611 (Check if this is true? OpenPGP calls them "strong".)
614 *) Merge the functionality of "dh" and "gendh" programs into a new program
615 "dhparam". The old programs are retained for now but will handle DH keys
616 (instead of parameters) in future.
619 *) Make the ciphers, s_server and s_client programs check the return values
620 when a new cipher list is set.
623 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
624 ciphers. Before when the 56bit ciphers were enabled the sorting was
627 The syntax for the cipher sorting has been extended to support sorting by
628 cipher-strength (using the strength_bits hard coded in the tables).
629 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
631 Fix a bug in the cipher-command parser: when supplying a cipher command
632 string with an "undefined" symbol (neither command nor alphanumeric
633 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
636 Due to the strength-sorting extension, the code of the
637 ssl_create_cipher_list() function was completely rearranged. I hope that
638 the readability was also increased :-)
639 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
641 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
642 for the first serial number and places 2 in the serial number file. This
643 avoids problems when the root CA is created with serial number zero and
644 the first user certificate has the same issuer name and serial number
648 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
649 the new code. Add documentation for this stuff.
652 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
653 X509_*() to X509at_*() on the grounds that they don't handle X509
654 structures and behave in an analagous way to the X509v3 functions:
655 they shouldn't be called directly but wrapper functions should be used
658 So we also now have some wrapper functions that call the X509at functions
659 when passed certificate requests. (TO DO: similar things can be done with
660 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
661 things. Some of these need some d2i or i2d and print functionality
662 because they handle more complex structures.)
665 *) Add missing #ifndefs that caused missing symbols when building libssl
666 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
668 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
670 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
671 has a return value which indicates the quality of the random data
672 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
673 error queue. New function RAND_pseudo_bytes() generates output that is
674 guaranteed to be unique but not unpredictable. RAND_add is like
675 RAND_seed, but takes an extra argument for an entropy estimate
676 (RAND_seed always assumes full entropy).
679 *) Do more iterations of Rabin-Miller probable prime test (specifically,
680 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
681 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
682 in crypto/bn/bn_prime.c for the complete table). This guarantees a
683 false-positive rate of at most 2^-80 for random input.
686 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
689 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
690 in the 0.9.5 release), this returns the chain
691 from an X509_CTX structure with a dup of the stack and all
692 the X509 reference counts upped: so the stack will exist
693 after X509_CTX_cleanup() has been called. Modify pkcs12.c
696 Also make SSL_SESSION_print() print out the verify return
700 *) Add manpage for the pkcs12 command. Also change the default
701 behaviour so MAC iteration counts are used unless the new
702 -nomaciter option is used. This improves file security and
703 only older versions of MSIE (4.0 for example) need it.
706 *) Honor the no-xxx Configure options when creating .DEF files.
709 *) Add PKCS#10 attributes to field table: challengePassword,
710 unstructuredName and unstructuredAddress. These are taken from
711 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
712 international characters are used.
714 More changes to X509_ATTRIBUTE code: allow the setting of types
715 based on strings. Remove the 'loc' parameter when adding
716 attributes because these will be a SET OF encoding which is sorted
720 *) Initial changes to the 'req' utility to allow request generation
721 automation. This will allow an application to just generate a template
722 file containing all the field values and have req construct the
725 Initial support for X509_ATTRIBUTE handling. Stacks of these are
726 used all over the place including certificate requests and PKCS#7
727 structures. They are currently handled manually where necessary with
728 some primitive wrappers for PKCS#7. The new functions behave in a
729 manner analogous to the X509 extension functions: they allow
730 attributes to be looked up by NID and added.
732 Later something similar to the X509V3 code would be desirable to
733 automatically handle the encoding, decoding and printing of the
734 more complex types. The string types like challengePassword can
735 be handled by the string table functions.
737 Also modified the multi byte string table handling. Now there is
738 a 'global mask' which masks out certain types. The table itself
739 can use the flag STABLE_NO_MASK to ignore the mask setting: this
740 is useful when for example there is only one permissible type
741 (as in countryName) and using the mask might result in no valid
745 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
746 SSL_get_peer_finished to allow applications to obtain the latest
747 Finished messages sent to the peer or expected from the peer,
748 respectively. (SSL_get_peer_finished is usually the Finished message
749 actually received from the peer, otherwise the protocol will be aborted.)
751 As the Finished message are message digests of the complete handshake
752 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
753 be used for external authentication procedures when the authentication
754 provided by SSL/TLS is not desired or is not enough.
757 *) Enhanced support for Alpha Linux is added. Now ./config checks if
758 the host supports BWX extension and if Compaq C is present on the
759 $PATH. Just exploiting of the BWX extension results in 20-30%
760 performance kick for some algorithms, e.g. DES and RC4 to mention
761 a couple. Compaq C in turn generates ~20% faster code for MD5 and
765 *) Add support for MS "fast SGC". This is arguably a violation of the
766 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
767 weak crypto and after checking the certificate is SGC a second one
768 with strong crypto. MS SGC stops the first handshake after receiving
769 the server certificate message and sends a second client hello. Since
770 a server will typically do all the time consuming operations before
771 expecting any further messages from the client (server key exchange
772 is the most expensive) there is little difference between the two.
774 To get OpenSSL to support MS SGC we have to permit a second client
775 hello message after we have sent server done. In addition we have to
776 reset the MAC if we do get this second client hello.
779 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
780 if a DER encoded private key is RSA or DSA traditional format. Changed
781 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
782 format DER encoded private key. Newer code should use PKCS#8 format which
783 has the key type encoded in the ASN1 structure. Added DER private key
784 support to pkcs8 application.
787 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
788 ciphersuites has been selected (as required by the SSL 3/TLS 1
789 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
790 is set, we interpret this as a request to violate the specification
791 (the worst that can happen is a handshake failure, and 'correct'
792 behaviour would result in a handshake failure anyway).
795 *) In SSL_CTX_add_session, take into account that there might be multiple
796 SSL_SESSION structures with the same session ID (e.g. when two threads
797 concurrently obtain them from an external cache).
798 The internal cache can handle only one SSL_SESSION with a given ID,
799 so if there's a conflict, we now throw out the old one to achieve
803 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
804 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
805 some routines that use cipher OIDs: some ciphers do not have OIDs
806 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
810 *) Simplify the trust setting structure and code. Now we just have
811 two sequences of OIDs for trusted and rejected settings. These will
812 typically have values the same as the extended key usage extension
813 and any application specific purposes.
815 The trust checking code now has a default behaviour: it will just
816 check for an object with the same NID as the passed id. Functions can
817 be provided to override either the default behaviour or the behaviour
818 for a given id. SSL client, server and email already have functions
819 in place for compatibility: they check the NID and also return "trusted"
820 if the certificate is self signed.
823 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
824 traditional format into an EVP_PKEY structure.
827 *) Add a password callback function PEM_cb() which either prompts for
828 a password if usr_data is NULL or otherwise assumes it is a null
829 terminated password. Allow passwords to be passed on command line
830 environment or config files in a few more utilities.
833 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
834 keys. Add some short names for PKCS#8 PBE algorithms and allow them
835 to be specified on the command line for the pkcs8 and pkcs12 utilities.
836 Update documentation.
839 *) Support for ASN1 "NULL" type. This could be handled before by using
840 ASN1_TYPE but there wasn't any function that would try to read a NULL
841 and produce an error if it couldn't. For compatibility we also have
842 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
843 don't allocate anything because they don't need to.
846 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
848 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
850 *) Rebuild of the memory allocation routines used by OpenSSL code and
851 possibly others as well. The purpose is to make an interface that
852 provide hooks so anyone can build a separate set of allocation and
853 deallocation routines to be used by OpenSSL, for example memory
854 pool implementations, or something else, which was previously hard
855 since Malloc(), Realloc() and Free() were defined as macros having
856 the values malloc, realloc and free, respectively (except for Win32
857 compilations). The same is provided for memory debugging code.
858 OpenSSL already comes with functionality to find memory leaks, but
859 this gives people a chance to debug other memory problems.
861 With these changes, a new set of functions and macros have appeared:
863 CRYPTO_set_mem_debug_functions() [F]
864 CRYPTO_get_mem_debug_functions() [F]
865 CRYPTO_dbg_set_options() [F]
866 CRYPTO_dbg_get_options() [F]
867 CRYPTO_malloc_debug_init() [M]
869 The memory debug functions are NULL by default, unless the library
870 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
871 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
872 gives the standard debugging functions that come with OpenSSL) or
873 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
874 provided by the library user) must be used. When the standard
875 debugging functions are used, CRYPTO_dbg_set_options can be used to
876 request additional information:
877 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
878 the CRYPTO_MDEBUG_xxx macro when compiling the library.
880 Also, things like CRYPTO_set_mem_functions will always give the
881 expected result (the new set of functions is used for allocation
882 and deallocation) at all times, regardless of platform and compiler
885 To finish it up, some functions that were never use in any other
886 way than through macros have a new API and new semantic:
892 All macros of value have retained their old syntax.
893 [Richard Levitte and Bodo Moeller]
895 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
896 ordering of SMIMECapabilities wasn't in "strength order" and there
897 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
901 *) Some ASN1 types with illegal zero length encoding (INTEGER,
902 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
903 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
905 *) Merge in my S/MIME library for OpenSSL. This provides a simple
906 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
907 functionality to handle multipart/signed properly) and a utility
908 called 'smime' to call all this stuff. This is based on code I
909 originally wrote for Celo who have kindly allowed it to be
913 *) Add variants des_set_key_checked and des_set_key_unchecked of
914 des_set_key (aka des_key_sched). Global variable des_check_key
915 decides which of these is called by des_set_key; this way
916 des_check_key behaves as it always did, but applications and
917 the library itself, which was buggy for des_check_key == 1,
918 have a cleaner way to pick the version they need.
921 *) New function PKCS12_newpass() which changes the password of a
925 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
926 dynamic mix. In both cases the ids can be used as an index into the
927 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
928 functions so they accept a list of the field values and the
929 application doesn't need to directly manipulate the X509_TRUST
933 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
937 *) Modify the way the V3 extension code looks up extensions. This now
938 works in a similar way to the object code: we have some "standard"
939 extensions in a static table which is searched with OBJ_bsearch()
940 and the application can add dynamic ones if needed. The file
941 crypto/x509v3/ext_dat.h now has the info: this file needs to be
942 updated whenever a new extension is added to the core code and kept
943 in ext_nid order. There is a simple program 'tabtest.c' which checks
944 this. New extensions are not added too often so this file can readily
945 be maintained manually.
947 There are two big advantages in doing things this way. The extensions
948 can be looked up immediately and no longer need to be "added" using
949 X509V3_add_standard_extensions(): this function now does nothing.
950 [Side note: I get *lots* of email saying the extension code doesn't
951 work because people forget to call this function]
952 Also no dynamic allocation is done unless new extensions are added:
953 so if we don't add custom extensions there is no need to call
954 X509V3_EXT_cleanup().
957 *) Modify enc utility's salting as follows: make salting the default. Add a
958 magic header, so unsalted files fail gracefully instead of just decrypting
959 to garbage. This is because not salting is a big security hole, so people
960 should be discouraged from doing it.
963 *) Fixes and enhancements to the 'x509' utility. It allowed a message
964 digest to be passed on the command line but it only used this
965 parameter when signing a certificate. Modified so all relevant
966 operations are affected by the digest parameter including the
967 -fingerprint and -x509toreq options. Also -x509toreq choked if a
968 DSA key was used because it didn't fix the digest.
971 *) Initial certificate chain verify code. Currently tests the untrusted
972 certificates for consistency with the verify purpose (which is set
973 when the X509_STORE_CTX structure is set up) and checks the pathlength.
975 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
976 this is because it will reject chains with invalid extensions whereas
977 every previous version of OpenSSL and SSLeay made no checks at all.
979 Trust code: checks the root CA for the relevant trust settings. Trust
980 settings have an initial value consistent with the verify purpose: e.g.
981 if the verify purpose is for SSL client use it expects the CA to be
982 trusted for SSL client use. However the default value can be changed to
983 permit custom trust settings: one example of this would be to only trust
984 certificates from a specific "secure" set of CAs.
986 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
987 which should be used for version portability: especially since the
988 verify structure is likely to change more often now.
990 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
991 to set them. If not set then assume SSL clients will verify SSL servers
994 Two new options to the verify program: -untrusted allows a set of
995 untrusted certificates to be passed in and -purpose which sets the
996 intended purpose of the certificate. If a purpose is set then the
997 new chain verify code is used to check extension consistency.
1000 *) Support for the authority information access extension.
1003 *) Modify RSA and DSA PEM read routines to transparently handle
1004 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
1005 public keys in a format compatible with certificate
1006 SubjectPublicKeyInfo structures. Unfortunately there were already
1007 functions called *_PublicKey_* which used various odd formats so
1008 these are retained for compatibility: however the DSA variants were
1009 never in a public release so they have been deleted. Changed dsa/rsa
1010 utilities to handle the new format: note no releases ever handled public
1011 keys so we should be OK.
1013 The primary motivation for this change is to avoid the same fiasco
1014 that dogs private keys: there are several incompatible private key
1015 formats some of which are standard and some OpenSSL specific and
1016 require various evil hacks to allow partial transparent handling and
1017 even then it doesn't work with DER formats. Given the option anything
1018 other than PKCS#8 should be dumped: but the other formats have to
1019 stay in the name of compatibility.
1021 With public keys and the benefit of hindsight one standard format
1022 is used which works with EVP_PKEY, RSA or DSA structures: though
1023 it clearly returns an error if you try to read the wrong kind of key.
1025 Added a -pubkey option to the 'x509' utility to output the public key.
1026 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
1027 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
1028 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
1029 that do the same as the EVP_PKEY_assign_*() except they up the
1030 reference count of the added key (they don't "swallow" the
1034 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
1035 CRLs would fail if the file contained no certificates or no CRLs:
1036 added a new function to read in both types and return the number
1037 read: this means that if none are read it will be an error. The
1038 DER versions of the certificate and CRL reader would always fail
1039 because it isn't possible to mix certificates and CRLs in DER format
1040 without choking one or the other routine. Changed this to just read
1041 a certificate: this is the best we can do. Also modified the code
1042 in apps/verify.c to take notice of return codes: it was previously
1043 attempting to read in certificates from NULL pointers and ignoring
1044 any errors: this is one reason why the cert and CRL reader seemed
1045 to work. It doesn't check return codes from the default certificate
1046 routines: these may well fail if the certificates aren't installed.
1049 *) Code to support otherName option in GeneralName.
1052 *) First update to verify code. Change the verify utility
1053 so it warns if it is passed a self signed certificate:
1054 for consistency with the normal behaviour. X509_verify
1055 has been modified to it will now verify a self signed
1056 certificate if *exactly* the same certificate appears
1057 in the store: it was previously impossible to trust a
1058 single self signed certificate. This means that:
1059 openssl verify ss.pem
1060 now gives a warning about a self signed certificate but
1061 openssl verify -CAfile ss.pem ss.pem
1065 *) For servers, store verify_result in SSL_SESSION data structure
1066 (and add it to external session representation).
1067 This is needed when client certificate verifications fails,
1068 but an application-provided verification callback (set by
1069 SSL_CTX_set_cert_verify_callback) allows accepting the session
1070 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
1071 but returns 1): When the session is reused, we have to set
1072 ssl->verify_result to the appropriate error code to avoid
1074 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
1076 *) Fix a bug in the new PKCS#7 code: it didn't consider the
1077 case in PKCS7_dataInit() where the signed PKCS7 structure
1078 didn't contain any existing data because it was being created.
1079 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
1081 *) Add a salt to the key derivation routines in enc.c. This
1082 forms the first 8 bytes of the encrypted file. Also add a
1083 -S option to allow a salt to be input on the command line.
1086 *) New function X509_cmp(). Oddly enough there wasn't a function
1087 to compare two certificates. We do this by working out the SHA1
1088 hash and comparing that. X509_cmp() will be needed by the trust
1092 *) SSL_get1_session() is like SSL_get_session(), but increments
1093 the reference count in the SSL_SESSION returned.
1094 [Geoff Thorpe <geoff@eu.c2.net>]
1096 *) Fix for 'req': it was adding a null to request attributes.
1097 Also change the X509_LOOKUP and X509_INFO code to handle
1098 certificate auxiliary information.
1101 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
1105 *) Add the possibility to add extra information to the memory leak
1106 detecting output, to form tracebacks, showing from where each
1107 allocation was originated: CRYPTO_push_info("constant string") adds
1108 the string plus current file name and line number to a per-thread
1109 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
1110 is like calling CYRPTO_pop_info() until the stack is empty.
1111 Also updated memory leak detection code to be multi-thread-safe.
1114 *) Add options -text and -noout to pkcs7 utility and delete the
1115 encryption options which never did anything. Update docs.
1118 *) Add options to some of the utilities to allow the pass phrase
1119 to be included on either the command line (not recommended on
1120 OSes like Unix) or read from the environment. Update the
1121 manpages and fix a few bugs.
1124 *) Add a few manpages for some of the openssl commands.
1127 *) Fix the -revoke option in ca. It was freeing up memory twice,
1128 leaking and not finding already revoked certificates.
1131 *) Extensive changes to support certificate auxiliary information.
1132 This involves the use of X509_CERT_AUX structure and X509_AUX
1133 functions. An X509_AUX function such as PEM_read_X509_AUX()
1134 can still read in a certificate file in the usual way but it
1135 will also read in any additional "auxiliary information". By
1136 doing things this way a fair degree of compatibility can be
1137 retained: existing certificates can have this information added
1138 using the new 'x509' options.
1140 Current auxiliary information includes an "alias" and some trust
1141 settings. The trust settings will ultimately be used in enhanced
1142 certificate chain verification routines: currently a certificate
1143 can only be trusted if it is self signed and then it is trusted
1147 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
1148 The problem was that one of the replacement routines had not been working
1149 since SSLeay releases. For now the offending routine has been replaced
1150 with non-optimised assembler. Even so, this now gives around 95%
1151 performance improvement for 1024 bit RSA signs.
1154 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
1155 handling. Most clients have the effective key size in bits equal to
1156 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
1157 A few however don't do this and instead use the size of the decrypted key
1158 to determine the RC2 key length and the AlgorithmIdentifier to determine
1159 the effective key length. In this case the effective key length can still
1160 be 40 bits but the key length can be 168 bits for example. This is fixed
1161 by manually forcing an RC2 key into the EVP_PKEY structure because the
1162 EVP code can't currently handle unusual RC2 key sizes: it always assumes
1163 the key length and effective key length are equal.
1166 *) Add a bunch of functions that should simplify the creation of
1167 X509_NAME structures. Now you should be able to do:
1168 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
1169 and have it automatically work out the correct field type and fill in
1170 the structures. The more adventurous can try:
1171 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
1172 and it will (hopefully) work out the correct multibyte encoding.
1175 *) Change the 'req' utility to use the new field handling and multibyte
1176 copy routines. Before the DN field creation was handled in an ad hoc
1177 way in req, ca, and x509 which was rather broken and didn't support
1178 BMPStrings or UTF8Strings. Since some software doesn't implement
1179 BMPStrings or UTF8Strings yet, they can be enabled using the config file
1180 using the dirstring_type option. See the new comment in the default
1181 openssl.cnf for more info.
1184 *) Make crypto/rand/md_rand.c more robust:
1185 - Assure unique random numbers after fork().
1186 - Make sure that concurrent threads access the global counter and
1187 md serializably so that we never lose entropy in them
1188 or use exactly the same state in multiple threads.
1189 Access to the large state is not always serializable because
1190 the additional locking could be a performance killer, and
1191 md should be large enough anyway.
1194 *) New file apps/app_rand.c with commonly needed functionality
1195 for handling the random seed file.
1197 Use the random seed file in some applications that previously did not:
1199 dsaparam -genkey (which also ignored its '-rand' option),
1202 x509 (when signing).
1203 Except on systems with /dev/urandom, it is crucial to have a random
1204 seed file at least for key creation, DSA signing, and for DH exchanges;
1205 for RSA signatures we could do without one.
1207 gendh and gendsa (unlike genrsa) used to read only the first byte
1208 of each file listed in the '-rand' option. The function as previously
1209 found in genrsa is now in app_rand.c and is used by all programs
1210 that support '-rand'.
1213 *) In RAND_write_file, use mode 0600 for creating files;
1214 don't just chmod when it may be too late.
1217 *) Report an error from X509_STORE_load_locations
1218 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
1221 *) New function ASN1_mbstring_copy() this copies a string in either
1222 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
1223 into an ASN1_STRING type. A mask of permissible types is passed
1224 and it chooses the "minimal" type to use or an error if not type
1228 *) Add function equivalents to the various macros in asn1.h. The old
1229 macros are retained with an M_ prefix. Code inside the library can
1230 use the M_ macros. External code (including the openssl utility)
1231 should *NOT* in order to be "shared library friendly".
1234 *) Add various functions that can check a certificate's extensions
1235 to see if it usable for various purposes such as SSL client,
1236 server or S/MIME and CAs of these types. This is currently
1237 VERY EXPERIMENTAL but will ultimately be used for certificate chain
1238 verification. Also added a -purpose flag to x509 utility to
1239 print out all the purposes.
1242 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
1246 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
1247 for, obtain and decode and extension and obtain its critical flag.
1248 This allows all the necessary extension code to be handled in a
1249 single function call.
1252 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
1253 platforms. See crypto/rc4/rc4_enc.c for further details.
1256 *) New -noout option to asn1parse. This causes no output to be produced
1257 its main use is when combined with -strparse and -out to extract data
1258 from a file (which may not be in ASN.1 format).
1261 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
1262 when producing the local key id.
1263 [Richard Levitte <levitte@stacken.kth.se>]
1265 *) New option -dhparam in s_server. This allows a DH parameter file to be
1266 stated explicitly. If it is not stated then it tries the first server
1267 certificate file. The previous behaviour hard coded the filename
1271 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
1272 a public key to be input or output. For example:
1273 openssl rsa -in key.pem -pubout -out pubkey.pem
1274 Also added necessary DSA public key functions to handle this.
1277 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
1278 in the message. This was handled by allowing
1279 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
1280 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
1282 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
1283 to the end of the strings whereas this didn't. This would cause problems
1284 if strings read with d2i_ASN1_bytes() were later modified.
1285 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
1287 *) Fix for base64 decode bug. When a base64 bio reads only one line of
1288 data and it contains EOF it will end up returning an error. This is
1289 caused by input 46 bytes long. The cause is due to the way base64
1290 BIOs find the start of base64 encoded data. They do this by trying a
1291 trial decode on each line until they find one that works. When they
1292 do a flag is set and it starts again knowing it can pass all the
1293 data directly through the decoder. Unfortunately it doesn't reset
1294 the context it uses. This means that if EOF is reached an attempt
1295 is made to pass two EOFs through the context and this causes the
1296 resulting error. This can also cause other problems as well. As is
1297 usual with these problems it takes *ages* to find and the fix is
1298 trivial: move one line.
1299 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
1301 *) Ugly workaround to get s_client and s_server working under Windows. The
1302 old code wouldn't work because it needed to select() on sockets and the
1303 tty (for keypresses and to see if data could be written). Win32 only
1304 supports select() on sockets so we select() with a 1s timeout on the
1305 sockets and then see if any characters are waiting to be read, if none
1306 are present then we retry, we also assume we can always write data to
1307 the tty. This isn't nice because the code then blocks until we've
1308 received a complete line of data and it is effectively polling the
1309 keyboard at 1s intervals: however it's quite a bit better than not
1310 working at all :-) A dedicated Windows application might handle this
1311 with an event loop for example.
1314 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
1315 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
1316 will be called when RSA_sign() and RSA_verify() are used. This is useful
1317 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
1318 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
1319 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
1320 This necessitated the support of an extra signature type NID_md5_sha1
1321 for SSL signatures and modifications to the SSL library to use it instead
1322 of calling RSA_public_decrypt() and RSA_private_encrypt().
1325 *) Add new -verify -CAfile and -CApath options to the crl program, these
1326 will lookup a CRL issuers certificate and verify the signature in a
1327 similar way to the verify program. Tidy up the crl program so it
1328 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
1329 less strict. It will now permit CRL extensions even if it is not
1330 a V2 CRL: this will allow it to tolerate some broken CRLs.
1333 *) Initialize all non-automatic variables each time one of the openssl
1334 sub-programs is started (this is necessary as they may be started
1335 multiple times from the "OpenSSL>" prompt).
1336 [Lennart Bang, Bodo Moeller]
1338 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
1339 removing all other RSA functionality (this is what NO_RSA does). This
1340 is so (for example) those in the US can disable those operations covered
1341 by the RSA patent while allowing storage and parsing of RSA keys and RSA
1345 *) Non-copying interface to BIO pairs.
1346 (still largely untested)
1349 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
1350 ASCII string. This was handled independently in various places before.
1353 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
1354 UTF8 strings a character at a time.
1357 *) Use client_version from client hello to select the protocol
1358 (s23_srvr.c) and for RSA client key exchange verification
1359 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
1362 *) Add various utility functions to handle SPKACs, these were previously
1363 handled by poking round in the structure internals. Added new function
1364 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
1365 print, verify and generate SPKACs. Based on an original idea from
1366 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
1369 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
1372 *) Allow the config file extension section to be overwritten on the
1373 command line. Based on an original idea from Massimiliano Pala
1374 <madwolf@comune.modena.it>. The new option is called -extensions
1375 and can be applied to ca, req and x509. Also -reqexts to override
1376 the request extensions in req and -crlexts to override the crl extensions
1380 *) Add new feature to the SPKAC handling in ca. Now you can include
1381 the same field multiple times by preceding it by "XXXX." for example:
1384 this is the same syntax as used in the req config file.
1387 *) Allow certificate extensions to be added to certificate requests. These
1388 are specified in a 'req_extensions' option of the req section of the
1389 config file. They can be printed out with the -text option to req but
1390 are otherwise ignored at present.
1393 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
1394 data read consists of only the final block it would not decrypted because
1395 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
1396 A misplaced 'break' also meant the decrypted final block might not be
1397 copied until the next read.
1400 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
1401 a few extra parameters to the DH structure: these will be useful if
1402 for example we want the value of 'q' or implement X9.42 DH.
1405 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
1406 provides hooks that allow the default DSA functions or functions on a
1407 "per key" basis to be replaced. This allows hardware acceleration and
1408 hardware key storage to be handled without major modification to the
1409 library. Also added low level modexp hooks and CRYPTO_EX structure and
1410 associated functions.
1413 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
1414 as "read only": it can't be written to and the buffer it points to will
1415 not be freed. Reading from a read only BIO is much more efficient than
1416 a normal memory BIO. This was added because there are several times when
1417 an area of memory needs to be read from a BIO. The previous method was
1418 to create a memory BIO and write the data to it, this results in two
1419 copies of the data and an O(n^2) reading algorithm. There is a new
1420 function BIO_new_mem_buf() which creates a read only memory BIO from
1421 an area of memory. Also modified the PKCS#7 routines to use read only
1425 *) Bugfix: ssl23_get_client_hello did not work properly when called in
1426 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
1427 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
1428 but a retry condition occured while trying to read the rest.
1431 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
1432 NID_pkcs7_encrypted by default: this was wrong since this should almost
1433 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
1434 the encrypted data type: this is a more sensible place to put it and it
1435 allows the PKCS#12 code to be tidied up that duplicated this
1439 *) Changed obj_dat.pl script so it takes its input and output files on
1440 the command line. This should avoid shell escape redirection problems
1444 *) Initial support for certificate extension requests, these are included
1445 in things like Xenroll certificate requests. Included functions to allow
1446 extensions to be obtained and added.
1449 *) -crlf option to s_client and s_server for sending newlines as
1450 CRLF (as required by many protocols).
1453 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
1455 *) Install libRSAglue.a when OpenSSL is built with RSAref.
1456 [Ralf S. Engelschall]
1458 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
1459 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
1461 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
1465 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
1466 DH parameters/keys (q is lost during that conversion, but the resulting
1467 DH parameters contain its length).
1469 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
1470 much faster than DH_generate_parameters (which creates parameters
1471 where p = 2*q + 1), and also the smaller q makes DH computations
1472 much more efficient (160-bit exponentiation instead of 1024-bit
1473 exponentiation); so this provides a convenient way to support DHE
1474 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
1475 utter importance to use
1476 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
1478 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
1479 when such DH parameters are used, because otherwise small subgroup
1480 attacks may become possible!
1483 *) Avoid memory leak in i2d_DHparams.
1486 *) Allow the -k option to be used more than once in the enc program:
1487 this allows the same encrypted message to be read by multiple recipients.
1490 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
1491 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
1492 it will always use the numerical form of the OID, even if it has a short
1496 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
1497 method only got called if p,q,dmp1,dmq1,iqmp components were present,
1498 otherwise bn_mod_exp was called. In the case of hardware keys for example
1499 no private key components need be present and it might store extra data
1500 in the RSA structure, which cannot be accessed from bn_mod_exp.
1501 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
1502 private key operations.
1505 *) Added support for SPARC Linux.
1508 *) pem_password_cb function type incompatibly changed from
1509 typedef int pem_password_cb(char *buf, int size, int rwflag);
1511 ....(char *buf, int size, int rwflag, void *userdata);
1512 so that applications can pass data to their callbacks:
1513 The PEM[_ASN1]_{read,write}... functions and macros now take an
1514 additional void * argument, which is just handed through whenever
1515 the password callback is called.
1516 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
1518 New function SSL_CTX_set_default_passwd_cb_userdata.
1520 Compatibility note: As many C implementations push function arguments
1521 onto the stack in reverse order, the new library version is likely to
1522 interoperate with programs that have been compiled with the old
1523 pem_password_cb definition (PEM_whatever takes some data that
1524 happens to be on the stack as its last argument, and the callback
1525 just ignores this garbage); but there is no guarantee whatsoever that
1528 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
1529 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
1530 problems not only on Windows, but also on some Unix platforms.
1531 To avoid problematic command lines, these definitions are now in an
1532 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
1533 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
1536 *) MIPS III/IV assembler module is reimplemented.
1539 *) More DES library cleanups: remove references to srand/rand and
1540 delete an unused file.
1543 *) Add support for the the free Netwide assembler (NASM) under Win32,
1544 since not many people have MASM (ml) and it can be hard to obtain.
1545 This is currently experimental but it seems to work OK and pass all
1546 the tests. Check out INSTALL.W32 for info.
1549 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
1550 without temporary keys kept an extra copy of the server key,
1551 and connections with temporary keys did not free everything in case
1555 *) New function RSA_check_key and new openssl rsa option -check
1556 for verifying the consistency of RSA keys.
1557 [Ulf Moeller, Bodo Moeller]
1559 *) Various changes to make Win32 compile work:
1560 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
1561 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
1562 comparison" warnings.
1563 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
1566 *) Add a debugging option to PKCS#5 v2 key generation function: when
1567 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
1568 derived keys are printed to stderr.
1571 *) Copy the flags in ASN1_STRING_dup().
1572 [Roman E. Pavlov <pre@mo.msk.ru>]
1574 *) The x509 application mishandled signing requests containing DSA
1575 keys when the signing key was also DSA and the parameters didn't match.
1577 It was supposed to omit the parameters when they matched the signing key:
1578 the verifying software was then supposed to automatically use the CA's
1579 parameters if they were absent from the end user certificate.
1581 Omitting parameters is no longer recommended. The test was also
1582 the wrong way round! This was probably due to unusual behaviour in
1583 EVP_cmp_parameters() which returns 1 if the parameters match.
1584 This meant that parameters were omitted when they *didn't* match and
1585 the certificate was useless. Certificates signed with 'ca' didn't have
1587 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
1589 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
1590 The interface is as follows:
1591 Applications can use
1592 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
1593 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
1594 "off" is now the default.
1595 The library internally uses
1596 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
1597 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
1598 to disable memory-checking temporarily.
1600 Some inconsistent states that previously were possible (and were
1601 even the default) are now avoided.
1603 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
1604 with each memory chunk allocated; this is occasionally more helpful
1605 than just having a counter.
1607 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
1609 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
1613 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
1614 which largely parallels "options", but is for changing API behaviour,
1615 whereas "options" are about protocol behaviour.
1616 Initial "mode" flags are:
1618 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
1619 a single record has been written.
1620 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
1621 retries use the same buffer location.
1622 (But all of the contents must be
1626 *) Bugfix: SSL_set_mode ignored its parameter, only SSL_CTX_set_mode
1629 *) Fix problems with no-hmac etc.
1630 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
1632 *) New functions RSA_get_default_method(), RSA_set_method() and
1633 RSA_get_method(). These allows replacement of RSA_METHODs without having
1634 to mess around with the internals of an RSA structure.
1637 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
1638 Also really enable memory leak checks in openssl.c and in some
1640 [Chad C. Mulligan, Bodo Moeller]
1642 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
1643 up the length of negative integers. This has now been simplified to just
1644 store the length when it is first determined and use it later, rather
1645 than trying to keep track of where data is copied and updating it to
1647 [Steve Henson, reported by Brien Wheeler
1648 <bwheeler@authentica-security.com>]
1650 *) Add a new function PKCS7_signatureVerify. This allows the verification
1651 of a PKCS#7 signature but with the signing certificate passed to the
1652 function itself. This contrasts with PKCS7_dataVerify which assumes the
1653 certificate is present in the PKCS#7 structure. This isn't always the
1654 case: certificates can be omitted from a PKCS#7 structure and be
1655 distributed by "out of band" means (such as a certificate database).
1658 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
1659 function prototypes in pem.h, also change util/mkdef.pl to add the
1660 necessary function names.
1663 *) mk1mf.pl (used by Windows builds) did not properly read the
1664 options set by Configure in the top level Makefile, and Configure
1665 was not even able to write more than one option correctly.
1666 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
1669 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
1670 file to be loaded from a BIO or FILE pointer. The BIO version will
1671 for example allow memory BIOs to contain config info.
1674 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
1675 Whoever hopes to achieve shared-library compatibility across versions
1676 must use this, not the compile-time macro.
1677 (Exercise 0.9.4: Which is the minimum library version required by
1679 Note: All this applies only to multi-threaded programs, others don't
1683 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
1684 through a BIO pair triggered the default case, i.e.
1685 SSLerr(...,SSL_R_UNKNOWN_STATE).
1688 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
1689 can use the SSL library even if none of the specific BIOs is
1693 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
1694 for the encoded length.
1695 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
1697 *) Add initial documentation of the X509V3 functions.
1700 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
1701 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
1702 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
1703 secure PKCS#8 private key format with a high iteration count.
1706 *) Fix determination of Perl interpreter: A perl or perl5
1707 _directory_ in $PATH was also accepted as the interpreter.
1708 [Ralf S. Engelschall]
1710 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
1711 wrong with it but it was very old and did things like calling
1712 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
1716 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
1717 to use the new extension code.
1720 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
1721 with macros. This should make it easier to change their form, add extra
1722 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
1726 *) Add to configuration table a new entry that can specify an alternative
1727 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
1728 according to Mark Crispin <MRC@Panda.COM>.
1732 *) DES CBC did not update the IV. Weird.
1735 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
1736 Changing the behaviour of the former might break existing programs --
1737 where IV updating is needed, des_ncbc_encrypt can be used.
1740 *) When bntest is run from "make test" it drives bc to check its
1741 calculations, as well as internally checking them. If an internal check
1742 fails, it needs to cause bc to give a non-zero result or make test carries
1743 on without noticing the failure. Fixed.
1746 *) DES library cleanups.
1749 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
1750 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
1751 ciphers. NOTE: although the key derivation function has been verified
1752 against some published test vectors it has not been extensively tested
1753 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
1757 *) Instead of "mkdir -p", which is not fully portable, use new
1758 Perl script "util/mkdir-p.pl".
1761 *) Rewrite the way password based encryption (PBE) is handled. It used to
1762 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
1763 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
1764 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
1765 the 'parameter' field of the AlgorithmIdentifier is passed to the
1766 underlying key generation function so it must do its own ASN1 parsing.
1767 This has also changed the EVP_PBE_CipherInit() function which now has a
1768 'parameter' argument instead of literal salt and iteration count values
1769 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
1772 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
1773 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
1774 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
1775 KEY" because this clashed with PKCS#8 unencrypted string. Since this
1776 value was just used as a "magic string" and not used directly its
1777 value doesn't matter.
1780 *) Introduce some semblance of const correctness to BN. Shame C doesn't
1784 *) "linux-sparc64" configuration (ultrapenguin).
1785 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
1786 "linux-sparc" configuration.
1787 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
1789 *) config now generates no-xxx options for missing ciphers.
1792 *) Support the EBCDIC character set (work in progress).
1793 File ebcdic.c not yet included because it has a different license.
1794 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
1796 *) Support BS2000/OSD-POSIX.
1797 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
1799 *) Make callbacks for key generation use void * instead of char *.
1802 *) Make S/MIME samples compile (not yet tested).
1805 *) Additional typesafe stacks.
1808 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
1812 Changes between 0.9.3 and 0.9.3a [29 May 1999]
1814 *) New configuration variant "sco5-gcc".
1816 *) Updated some demos.
1817 [Sean O Riordain, Wade Scholine]
1819 *) Add missing BIO_free at exit of pkcs12 application.
1822 *) Fix memory leak in conf.c.
1825 *) Updates for Win32 to assembler version of MD5.
1828 *) Set #! path to perl in apps/der_chop to where we found it
1829 instead of using a fixed path.
1832 *) SHA library changes for irix64-mips4-cc.
1835 *) Improvements for VMS support.
1839 Changes between 0.9.2b and 0.9.3 [24 May 1999]
1841 *) Bignum library bug fix. IRIX 6 passes "make test" now!
1842 This also avoids the problems with SC4.2 and unpatched SC5.
1843 [Andy Polyakov <appro@fy.chalmers.se>]
1845 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
1846 These are required because of the typesafe stack would otherwise break
1847 existing code. If old code used a structure member which used to be STACK
1848 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
1849 sk_num or sk_value it would produce an error because the num, data members
1850 are not present in STACK_OF. Now it just produces a warning. sk_set
1851 replaces the old method of assigning a value to sk_value
1852 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
1853 that does this will no longer work (and should use sk_set instead) but
1854 this could be regarded as a "questionable" behaviour anyway.
1857 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
1858 correctly handle encrypted S/MIME data.
1861 *) Change type of various DES function arguments from des_cblock
1862 (which means, in function argument declarations, pointer to char)
1863 to des_cblock * (meaning pointer to array with 8 char elements),
1864 which allows the compiler to do more typechecking; it was like
1865 that back in SSLeay, but with lots of ugly casts.
1867 Introduce new type const_des_cblock.
1870 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
1871 problems: find RecipientInfo structure that matches recipient certificate
1872 and initialise the ASN1 structures properly based on passed cipher.
1875 *) Belatedly make the BN tests actually check the results.
1878 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
1879 to and from BNs: it was completely broken. New compilation option
1880 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
1881 key elements as negative integers.
1884 *) Reorganize and speed up MD5.
1885 [Andy Polyakov <appro@fy.chalmers.se>]
1888 [Richard Levitte <richard@levitte.org>]
1890 *) New option -out to asn1parse to allow the parsed structure to be
1891 output to a file. This is most useful when combined with the -strparse
1892 option to examine the output of things like OCTET STRINGS.
1895 *) Make SSL library a little more fool-proof by not requiring any longer
1896 that SSL_set_{accept,connect}_state be called before
1897 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
1898 in many applications because usually everything *appeared* to work as
1899 intended anyway -- now it really works as intended).
1902 *) Move openssl.cnf out of lib/.
1905 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
1906 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
1907 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
1908 [Ralf S. Engelschall]
1910 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
1911 handle PKCS#7 enveloped data properly.
1912 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
1914 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
1915 copying pointers. The cert_st handling is changed by this in
1916 various ways (and thus what used to be known as ctx->default_cert
1917 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
1918 any longer when s->cert does not give us what we need).
1919 ssl_cert_instantiate becomes obsolete by this change.
1920 As soon as we've got the new code right (possibly it already is?),
1921 we have solved a couple of bugs of the earlier code where s->cert
1922 was used as if it could not have been shared with other SSL structures.
1924 Note that using the SSL API in certain dirty ways now will result
1925 in different behaviour than observed with earlier library versions:
1926 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
1927 does not influence s as it used to.
1929 In order to clean up things more thoroughly, inside SSL_SESSION
1930 we don't use CERT any longer, but a new structure SESS_CERT
1931 that holds per-session data (if available); currently, this is
1932 the peer's certificate chain and, for clients, the server's certificate
1933 and temporary key. CERT holds only those values that can have
1934 meaningful defaults in an SSL_CTX.
1937 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
1938 from the internal representation. Various PKCS#7 fixes: remove some
1939 evil casts and set the enc_dig_alg field properly based on the signing
1943 *) Allow PKCS#12 password to be set from the command line or the
1944 environment. Let 'ca' get its config file name from the environment
1945 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
1949 *) Allow certificate policies extension to use an IA5STRING for the
1950 organization field. This is contrary to the PKIX definition but
1951 VeriSign uses it and IE5 only recognises this form. Document 'x509'
1955 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
1956 without disallowing inline assembler and the like for non-pedantic builds.
1959 *) Support Borland C++ builder.
1960 [Janez Jere <jj@void.si>, modified by Ulf Möller]
1965 *) SHA-1 cleanups and performance enhancements.
1966 [Andy Polyakov <appro@fy.chalmers.se>]
1968 *) Sparc v8plus assembler for the bignum library.
1969 [Andy Polyakov <appro@fy.chalmers.se>]
1971 *) Accept any -xxx and +xxx compiler options in Configure.
1974 *) Update HPUX configuration.
1977 *) Add missing sk_<type>_unshift() function to safestack.h
1978 [Ralf S. Engelschall]
1980 *) New function SSL_CTX_use_certificate_chain_file that sets the
1981 "extra_cert"s in addition to the certificate. (This makes sense
1982 only for "PEM" format files, as chains as a whole are not
1986 *) Support verify_depth from the SSL API.
1987 x509_vfy.c had what can be considered an off-by-one-error:
1988 Its depth (which was not part of the external interface)
1989 was actually counting the number of certificates in a chain;
1990 now it really counts the depth.
1993 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
1994 instead of X509err, which often resulted in confusing error
1995 messages since the error codes are not globally unique
1996 (e.g. an alleged error in ssl3_accept when a certificate
1997 didn't match the private key).
1999 *) New function SSL_CTX_set_session_id_context that allows to set a default
2000 value (so that you don't need SSL_set_session_id_context for each
2001 connection using the SSL_CTX).
2004 *) OAEP decoding bug fix.
2007 *) Support INSTALL_PREFIX for package builders, as proposed by
2011 *) New Configure options "threads" and "no-threads". For systems
2012 where the proper compiler options are known (currently Solaris
2013 and Linux), "threads" is the default.
2016 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
2019 *) Install various scripts to $(OPENSSLDIR)/misc, not to
2020 $(INSTALLTOP)/bin -- they shouldn't clutter directories
2021 such as /usr/local/bin.
2024 *) "make linux-shared" to build shared libraries.
2025 [Niels Poppe <niels@netbox.org>]
2027 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
2030 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
2031 extension adding in x509 utility.
2034 *) Remove NOPROTO sections and error code comments.
2037 *) Partial rewrite of the DEF file generator to now parse the ANSI
2041 *) New Configure options --prefix=DIR and --openssldir=DIR.
2044 *) Complete rewrite of the error code script(s). It is all now handled
2045 by one script at the top level which handles error code gathering,
2046 header rewriting and C source file generation. It should be much better
2047 than the old method: it now uses a modified version of Ulf's parser to
2048 read the ANSI prototypes in all header files (thus the old K&R definitions
2049 aren't needed for error creation any more) and do a better job of
2050 translating function codes into names. The old 'ASN1 error code imbedded
2051 in a comment' is no longer necessary and it doesn't use .err files which
2052 have now been deleted. Also the error code call doesn't have to appear all
2053 on one line (which resulted in some large lines...).
2056 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
2059 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
2060 0 (which usually indicates a closed connection), but continue reading.
2063 *) Fix some race conditions.
2066 *) Add support for CRL distribution points extension. Add Certificate
2067 Policies and CRL distribution points documentation.
2070 *) Move the autogenerated header file parts to crypto/opensslconf.h.
2073 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
2074 8 of keying material. Merlin has also confirmed interop with this fix
2075 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
2076 [Merlin Hughes <merlin@baltimore.ie>]
2078 *) Fix lots of warnings.
2079 [Richard Levitte <levitte@stacken.kth.se>]
2081 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
2082 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
2083 [Richard Levitte <levitte@stacken.kth.se>]
2085 *) Fix problems with sizeof(long) == 8.
2086 [Andy Polyakov <appro@fy.chalmers.se>]
2088 *) Change functions to ANSI C.
2091 *) Fix typos in error codes.
2092 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
2094 *) Remove defunct assembler files from Configure.
2097 *) SPARC v8 assembler BIGNUM implementation.
2098 [Andy Polyakov <appro@fy.chalmers.se>]
2100 *) Support for Certificate Policies extension: both print and set.
2101 Various additions to support the r2i method this uses.
2104 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
2105 return a const string when you are expecting an allocated buffer.
2108 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
2109 types DirectoryString and DisplayText.
2112 *) Add code to allow r2i extensions to access the configuration database,
2113 add an LHASH database driver and add several ctx helper functions.
2116 *) Fix an evil bug in bn_expand2() which caused various BN functions to
2117 fail when they extended the size of a BIGNUM.
2120 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
2121 support typesafe stack.
2124 *) Fix typo in SSL_[gs]et_options().
2125 [Nils Frostberg <nils@medcom.se>]
2127 *) Delete various functions and files that belonged to the (now obsolete)
2128 old X509V3 handling code.
2131 *) New Configure option "rsaref".
2134 *) Don't auto-generate pem.h.
2137 *) Introduce type-safe ASN.1 SETs.
2140 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
2141 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
2143 *) Introduce type-safe STACKs. This will almost certainly break lots of code
2144 that links with OpenSSL (well at least cause lots of warnings), but fear
2145 not: the conversion is trivial, and it eliminates loads of evil casts. A
2146 few STACKed things have been converted already. Feel free to convert more.
2147 In the fullness of time, I'll do away with the STACK type altogether.
2150 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
2151 specified in <certfile> by updating the entry in the index.txt file.
2152 This way one no longer has to edit the index.txt file manually for
2153 revoking a certificate. The -revoke option does the gory details now.
2154 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
2156 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
2157 `-text' option at all and this way the `-noout -text' combination was
2158 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
2159 [Ralf S. Engelschall]
2161 *) Make sure a corresponding plain text error message exists for the
2162 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
2163 verify callback function determined that a certificate was revoked.
2164 [Ralf S. Engelschall]
2166 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
2167 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
2168 all available cipers including rc5, which was forgotten until now.
2169 In order to let the testing shell script know which algorithms
2170 are available, a new (up to now undocumented) command
2171 "openssl list-cipher-commands" is used.
2174 *) Bugfix: s_client occasionally would sleep in select() when
2175 it should have checked SSL_pending() first.
2178 *) New functions DSA_do_sign and DSA_do_verify to provide access to
2179 the raw DSA values prior to ASN.1 encoding.
2182 *) Tweaks to Configure
2183 [Niels Poppe <niels@netbox.org>]
2185 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
2189 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
2192 *) New config option to avoid instructions that are illegal on the 80386.
2193 The default code is faster, but requires at least a 486.
2196 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
2197 SSL2_SERVER_VERSION (not used at all) macros, which are now the
2198 same as SSL2_VERSION anyway.
2201 *) New "-showcerts" option for s_client.
2204 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
2205 application. Various cleanups and fixes.
2208 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
2209 modify error routines to work internally. Add error codes and PBE init
2210 to library startup routines.
2213 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
2214 packing functions to asn1 and evp. Changed function names and error
2215 codes along the way.
2218 *) PKCS12 integration: and so it begins... First of several patches to
2219 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
2220 objects to objects.h
2223 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
2224 and display support for Thawte strong extranet extension.
2227 *) Add LinuxPPC support.
2228 [Jeff Dubrule <igor@pobox.org>]
2230 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
2231 bn_div_words in alpha.s.
2232 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
2234 *) Make sure the RSA OAEP test is skipped under -DRSAref because
2235 OAEP isn't supported when OpenSSL is built with RSAref.
2236 [Ulf Moeller <ulf@fitug.de>]
2238 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
2239 so they no longer are missing under -DNOPROTO.
2240 [Soren S. Jorvang <soren@t.dk>]
2243 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
2245 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
2246 doesn't work when the session is reused. Coming soon!
2249 *) Fix a security hole, that allows sessions to be reused in the wrong
2250 context thus bypassing client cert protection! All software that uses
2251 client certs and session caches in multiple contexts NEEDS PATCHING to
2252 allow session reuse! A fuller solution is in the works.
2253 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
2255 *) Some more source tree cleanups (removed obsolete files
2256 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
2257 permission on "config" script to be executable) and a fix for the INSTALL
2259 [Ulf Moeller <ulf@fitug.de>]
2261 *) Remove some legacy and erroneous uses of malloc, free instead of
2263 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
2265 *) Make rsa_oaep_test return non-zero on error.
2266 [Ulf Moeller <ulf@fitug.de>]
2268 *) Add support for native Solaris shared libraries. Configure
2269 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
2270 if someone would make that last step automatic.
2271 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
2273 *) ctx_size was not built with the right compiler during "make links". Fixed.
2276 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
2277 except NULL ciphers". This means the default cipher list will no longer
2278 enable NULL ciphers. They need to be specifically enabled e.g. with
2279 the string "DEFAULT:eNULL".
2282 *) Fix to RSA private encryption routines: if p < q then it would
2283 occasionally produce an invalid result. This will only happen with
2284 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
2287 *) Be less restrictive and allow also `perl util/perlpath.pl
2288 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
2289 because this way one can also use an interpreter named `perl5' (which is
2290 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
2291 installed as `perl').
2292 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
2294 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
2295 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
2297 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
2298 advapi32.lib to Win32 build and change the pem test comparision
2299 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
2300 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
2301 and crypto/des/ede_cbcm_enc.c.
2304 *) DES quad checksum was broken on big-endian architectures. Fixed.
2307 *) Comment out two functions in bio.h that aren't implemented. Fix up the
2308 Win32 test batch file so it (might) work again. The Win32 test batch file
2309 is horrible: I feel ill....
2312 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
2313 in e_os.h. Audit of header files to check ANSI and non ANSI
2314 sections: 10 functions were absent from non ANSI section and not exported
2315 from Windows DLLs. Fixed up libeay.num for new functions.
2318 *) Make `openssl version' output lines consistent.
2319 [Ralf S. Engelschall]
2321 *) Fix Win32 symbol export lists for BIO functions: Added
2322 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
2323 to ms/libeay{16,32}.def.
2324 [Ralf S. Engelschall]
2326 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
2327 fine under Unix and passes some trivial tests I've now added. But the
2328 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
2329 added to make sure no one expects that this stuff really works in the
2330 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
2331 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
2333 [Ralf S. Engelschall]
2335 *) Fix the generation of two part addresses in perl.
2336 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
2338 *) Add config entry for Linux on MIPS.
2339 [John Tobey <jtobey@channel1.com>]
2341 *) Make links whenever Configure is run, unless we are on Windoze.
2344 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
2345 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
2349 *) Add a useful kludge to allow package maintainers to specify compiler and
2350 other platforms details on the command line without having to patch the
2351 Configure script everytime: One now can use ``perl Configure
2352 <id>:<details>'', i.e. platform ids are allowed to have details appended
2353 to them (seperated by colons). This is treated as there would be a static
2354 pre-configured entry in Configure's %table under key <id> with value
2355 <details> and ``perl Configure <id>'' is called. So, when you want to
2356 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
2357 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
2358 now, which overrides the FreeBSD-elf entry on-the-fly.
2359 [Ralf S. Engelschall]
2361 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
2364 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
2365 on the `perl Configure ...' command line. This way one can compile
2366 OpenSSL libraries with Position Independent Code (PIC) which is needed
2367 for linking it into DSOs.
2368 [Ralf S. Engelschall]
2370 *) Remarkably, export ciphers were totally broken and no-one had noticed!
2374 *) Cleaned up the LICENSE document: The official contact for any license
2375 questions now is the OpenSSL core team under openssl-core@openssl.org.
2376 And add a paragraph about the dual-license situation to make sure people
2377 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
2378 to the OpenSSL toolkit.
2379 [Ralf S. Engelschall]
2381 *) General source tree makefile cleanups: Made `making xxx in yyy...'
2382 display consistent in the source tree and replaced `/bin/rm' by `rm'.
2383 Additonally cleaned up the `make links' target: Remove unnecessary
2384 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
2385 to speed processing and no longer clutter the display with confusing
2386 stuff. Instead only the actually done links are displayed.
2387 [Ralf S. Engelschall]
2389 *) Permit null encryption ciphersuites, used for authentication only. It used
2390 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
2391 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
2395 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
2396 signed attributes when verifying signatures (this would break them),
2397 the detached data encoding was wrong and public keys obtained using
2398 X509_get_pubkey() weren't freed.
2401 *) Add text documentation for the BUFFER functions. Also added a work around
2402 to a Win95 console bug. This was triggered by the password read stuff: the
2403 last character typed gets carried over to the next fread(). If you were
2404 generating a new cert request using 'req' for example then the last
2405 character of the passphrase would be CR which would then enter the first
2409 *) Added the new `Includes OpenSSL Cryptography Software' button as
2410 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
2411 button and can be used by applications based on OpenSSL to show the
2412 relationship to the OpenSSL project.
2413 [Ralf S. Engelschall]
2415 *) Remove confusing variables in function signatures in files
2416 ssl/ssl_lib.c and ssl/ssl.h.
2417 [Lennart Bong <lob@kulthea.stacken.kth.se>]
2419 *) Don't install bss_file.c under PREFIX/include/
2420 [Lennart Bong <lob@kulthea.stacken.kth.se>]
2422 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
2423 functions that return function pointers and has support for NT specific
2424 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
2425 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
2426 unsigned to signed types: this was killing the Win32 compile.
2429 *) Add new certificate file to stack functions,
2430 SSL_add_dir_cert_subjects_to_stack() and
2431 SSL_add_file_cert_subjects_to_stack(). These largely supplant
2432 SSL_load_client_CA_file(), and can be used to add multiple certs easily
2433 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
2434 This means that Apache-SSL and similar packages don't have to mess around
2435 to add as many CAs as they want to the preferred list.
2438 *) Experiment with doxygen documentation. Currently only partially applied to
2440 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
2441 openssl.doxy as the configuration file.
2444 *) Get rid of remaining C++-style comments which strict C compilers hate.
2445 [Ralf S. Engelschall, pointed out by Carlos Amengual]
2447 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
2448 compiled in by default: it has problems with large keys.
2451 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
2452 DH private keys and/or callback functions which directly correspond to
2453 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
2454 is needed for applications which have to configure certificates on a
2455 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
2457 For the RSA certificate situation is makes no difference, but
2458 for the DSA certificate situation this fixes the "no shared cipher"
2459 problem where the OpenSSL cipher selection procedure failed because the
2460 temporary keys were not overtaken from the context and the API provided
2461 no way to reconfigure them.
2462 The new functions now let applications reconfigure the stuff and they
2463 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
2464 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
2465 non-public-API function ssl_cert_instantiate() is used as a helper
2466 function and also to reduce code redundancy inside ssl_rsa.c.
2467 [Ralf S. Engelschall]
2469 *) Move s_server -dcert and -dkey options out of the undocumented feature
2470 area because they are useful for the DSA situation and should be
2471 recognized by the users.
2472 [Ralf S. Engelschall]
2474 *) Fix the cipher decision scheme for export ciphers: the export bits are
2475 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
2476 SSL_EXP_MASK. So, the original variable has to be used instead of the
2477 already masked variable.
2478 [Richard Levitte <levitte@stacken.kth.se>]
2480 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
2481 [Richard Levitte <levitte@stacken.kth.se>]
2483 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
2484 from `int' to `unsigned int' because it's a length and initialized by
2485 EVP_DigestFinal() which expects an `unsigned int *'.
2486 [Richard Levitte <levitte@stacken.kth.se>]
2488 *) Don't hard-code path to Perl interpreter on shebang line of Configure
2489 script. Instead use the usual Shell->Perl transition trick.
2490 [Ralf S. Engelschall]
2492 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
2493 (in addition to RSA certificates) to match the behaviour of `openssl dsa
2494 -noout -modulus' as it's already the case for `openssl rsa -noout
2495 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
2496 currently the public key is printed (a decision which was already done by
2497 `openssl dsa -modulus' in the past) which serves a similar purpose.
2498 Additionally the NO_RSA no longer completely removes the whole -modulus
2499 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
2501 [Ralf S. Engelschall]
2503 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
2504 BIO. See the source (crypto/evp/bio_ok.c) for more info.
2505 [Arne Ansper <arne@ats.cyber.ee>]
2507 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
2508 to be added. Now both 'req' and 'ca' can use new objects defined in the
2512 *) Add cool BIO that does syslog (or event log on NT).
2513 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
2515 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
2516 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
2517 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
2518 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
2521 *) Add preliminary config info for new extension code.
2524 *) Make RSA_NO_PADDING really use no padding.
2525 [Ulf Moeller <ulf@fitug.de>]
2527 *) Generate errors when private/public key check is done.
2530 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
2531 for some CRL extensions and new objects added.
2534 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
2535 key usage extension and fuller support for authority key id.
2538 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
2539 padding method for RSA, which is recommended for new applications in PKCS
2540 #1 v2.0 (RFC 2437, October 1998).
2541 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
2542 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
2543 against Bleichbacher's attack on RSA.
2544 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
2547 *) Updates to the new SSL compression code
2548 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
2550 *) Fix so that the version number in the master secret, when passed
2551 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
2552 (because the server will not accept higher), that the version number
2553 is 0x03,0x01, not 0x03,0x00
2554 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
2556 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
2557 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
2558 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
2561 *) Support for RAW extensions where an arbitrary extension can be
2562 created by including its DER encoding. See apps/openssl.cnf for
2566 *) Make sure latest Perl versions don't interpret some generated C array
2567 code as Perl array code in the crypto/err/err_genc.pl script.
2568 [Lars Weber <3weber@informatik.uni-hamburg.de>]
2570 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
2571 not many people have the assembler. Various Win32 compilation fixes and
2572 update to the INSTALL.W32 file with (hopefully) more accurate Win32
2576 *) Modify configure script 'Configure' to automatically create crypto/date.h
2577 file under Win32 and also build pem.h from pem.org. New script
2578 util/mkfiles.pl to create the MINFO file on environments that can't do a
2579 'make files': perl util/mkfiles.pl >MINFO should work.
2582 *) Major rework of DES function declarations, in the pursuit of correctness
2583 and purity. As a result, many evil casts evaporated, and some weirdness,
2584 too. You may find this causes warnings in your code. Zapping your evil
2585 casts will probably fix them. Mostly.
2588 *) Fix for a typo in asn1.h. Bug fix to object creation script
2589 obj_dat.pl. It considered a zero in an object definition to mean
2590 "end of object": none of the objects in objects.h have any zeros
2591 so it wasn't spotted.
2592 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
2594 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
2595 Masking (CBCM). In the absence of test vectors, the best I have been able
2596 to do is check that the decrypt undoes the encrypt, so far. Send me test
2597 vectors if you have them.
2600 *) Correct calculation of key length for export ciphers (too much space was
2601 allocated for null ciphers). This has not been tested!
2604 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
2605 message is now correct (it understands "crypto" and "ssl" on its
2606 command line). There is also now an "update" option. This will update
2607 the util/ssleay.num and util/libeay.num files with any new functions.
2609 perl util/mkdef.pl crypto ssl update
2610 it will update them.
2613 *) Overhauled the Perl interface (perl/*):
2614 - ported BN stuff to OpenSSL's different BN library
2615 - made the perl/ source tree CVS-aware
2616 - renamed the package from SSLeay to OpenSSL (the files still contain
2617 their history because I've copied them in the repository)
2618 - removed obsolete files (the test scripts will be replaced
2619 by better Test::Harness variants in the future)
2620 [Ralf S. Engelschall]
2622 *) First cut for a very conservative source tree cleanup:
2623 1. merge various obsolete readme texts into doc/ssleay.txt
2624 where we collect the old documents and readme texts.
2625 2. remove the first part of files where I'm already sure that we no
2626 longer need them because of three reasons: either they are just temporary
2627 files which were left by Eric or they are preserved original files where
2628 I've verified that the diff is also available in the CVS via "cvs diff
2629 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
2630 the crypto/md/ stuff).
2631 [Ralf S. Engelschall]
2633 *) More extension code. Incomplete support for subject and issuer alt
2634 name, issuer and authority key id. Change the i2v function parameters
2635 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
2636 what that's for :-) Fix to ASN1 macro which messed up
2637 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
2640 *) Preliminary support for ENUMERATED type. This is largely copied from the
2644 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
2645 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
2647 *) Make sure `make rehash' target really finds the `openssl' program.
2648 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
2650 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
2651 like to hear about it if this slows down other processors.
2654 *) Add CygWin32 platform information to Configure script.
2655 [Alan Batie <batie@aahz.jf.intel.com>]
2657 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
2658 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
2660 *) New program nseq to manipulate netscape certificate sequences
2663 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
2667 *) Fixes to BN code. Previously the default was to define BN_RECURSION
2668 but the BN code had some problems that would cause failures when
2669 doing certificate verification and some other functions.
2670 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
2672 *) Add ASN1 and PEM code to support netscape certificate sequences.
2675 *) Add ASN1 and PEM code to support netscape certificate sequences.
2678 *) Add several PKIX and private extended key usage OIDs.
2681 *) Modify the 'ca' program to handle the new extension code. Modify
2682 openssl.cnf for new extension format, add comments.
2685 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
2686 and add a sample to openssl.cnf so req -x509 now adds appropriate
2690 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
2691 error code, add initial support to X509_print() and x509 application.
2694 *) Takes a deep breath and start addding X509 V3 extension support code. Add
2695 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
2696 stuff is currently isolated and isn't even compiled yet.
2699 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
2700 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
2701 Removed the versions check from X509 routines when loading extensions:
2702 this allows certain broken certificates that don't set the version
2703 properly to be processed.
2706 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
2707 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
2708 can still be regenerated with "make depend".
2711 *) Spelling mistake in C version of CAST-128.
2712 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
2714 *) Changes to the error generation code. The perl script err-code.pl
2715 now reads in the old error codes and retains the old numbers, only
2716 adding new ones if necessary. It also only changes the .err files if new
2717 codes are added. The makefiles have been modified to only insert errors
2718 when needed (to avoid needlessly modifying header files). This is done
2719 by only inserting errors if the .err file is newer than the auto generated
2720 C file. To rebuild all the error codes from scratch (the old behaviour)
2721 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
2722 or delete all the .err files.
2725 *) CAST-128 was incorrectly implemented for short keys. The C version has
2726 been fixed, but is untested. The assembler versions are also fixed, but
2727 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
2728 to regenerate it if needed.
2729 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
2730 Hagino <itojun@kame.net>]
2732 *) File was opened incorrectly in randfile.c.
2733 [Ulf Möller <ulf@fitug.de>]
2735 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
2736 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
2737 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
2738 al: it's just almost always a UTCTime. Note this patch adds new error
2739 codes so do a "make errors" if there are problems.
2742 *) Correct Linux 1 recognition in config.
2743 [Ulf Möller <ulf@fitug.de>]
2745 *) Remove pointless MD5 hash when using DSA keys in ca.
2746 [Anonymous <nobody@replay.com>]
2748 *) Generate an error if given an empty string as a cert directory. Also
2749 generate an error if handed NULL (previously returned 0 to indicate an
2750 error, but didn't set one).
2751 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
2753 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
2756 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
2757 parameters. This was causing a warning which killed off the Win32 compile.
2760 *) Remove C++ style comments from crypto/bn/bn_local.h.
2761 [Neil Costigan <neil.costigan@celocom.com>]
2763 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
2764 based on a text string, looking up short and long names and finally
2765 "dot" format. The "dot" format stuff didn't work. Added new function
2766 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
2767 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
2768 OID is not part of the table.
2771 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
2772 X509_LOOKUP_by_alias().
2775 *) Sort openssl functions by name.
2778 *) Get the gendsa program working (hopefully) and add it to app list. Remove
2779 encryption from sample DSA keys (in case anyone is interested the password
2783 *) Make _all_ *_free functions accept a NULL pointer.
2784 [Frans Heymans <fheymans@isaserver.be>]
2786 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
2788 [Anonymous <nobody@replay.com>]
2790 *) s_server should send the CAfile as acceptable CAs, not its own cert.
2791 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
2793 *) Don't blow it for numeric -newkey arguments to apps/req.
2794 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
2796 *) Temp key "for export" tests were wrong in s3_srvr.c.
2797 [Anonymous <nobody@replay.com>]
2799 *) Add prototype for temp key callback functions
2800 SSL_CTX_set_tmp_{rsa,dh}_callback().
2803 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
2804 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
2807 *) X509_name_add_entry() freed the wrong thing after an error.
2808 [Arne Ansper <arne@ats.cyber.ee>]
2810 *) rsa_eay.c would attempt to free a NULL context.
2811 [Arne Ansper <arne@ats.cyber.ee>]
2813 *) BIO_s_socket() had a broken should_retry() on Windoze.
2814 [Arne Ansper <arne@ats.cyber.ee>]
2816 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
2817 [Arne Ansper <arne@ats.cyber.ee>]
2819 *) Make sure the already existing X509_STORE->depth variable is initialized
2820 in X509_STORE_new(), but document the fact that this variable is still
2821 unused in the certificate verification process.
2822 [Ralf S. Engelschall]
2824 *) Fix the various library and apps files to free up pkeys obtained from
2825 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
2828 *) Fix reference counting in X509_PUBKEY_get(). This makes
2829 demos/maurice/example2.c work, amongst others, probably.
2830 [Steve Henson and Ben Laurie]
2832 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
2833 `openssl' and second, the shortcut symlinks for the `openssl <command>'
2834 are no longer created. This way we have a single and consistent command
2835 line interface `openssl <command>', similar to `cvs <command>'.
2836 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
2838 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
2839 BIT STRING wrapper always have zero unused bits.
2842 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
2845 *) Make the top-level INSTALL documentation easier to understand.
2848 *) Makefiles updated to exit if an error occurs in a sub-directory
2849 make (including if user presses ^C) [Paul Sutton]
2851 *) Make Montgomery context stuff explicit in RSA data structure.
2854 *) Fix build order of pem and err to allow for generated pem.h.
2857 *) Fix renumbering bug in X509_NAME_delete_entry().
2860 *) Enhanced the err-ins.pl script so it makes the error library number
2861 global and can add a library name. This is needed for external ASN1 and
2862 other error libraries.
2865 *) Fixed sk_insert which never worked properly.
2868 *) Fix ASN1 macros so they can handle indefinite length construted
2869 EXPLICIT tags. Some non standard certificates use these: they can now
2873 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
2874 into a single doc/ssleay.txt bundle. This way the information is still
2875 preserved but no longer messes up this directory. Now it's new room for
2876 the new set of documenation files.
2877 [Ralf S. Engelschall]
2879 *) SETs were incorrectly DER encoded. This was a major pain, because they
2880 shared code with SEQUENCEs, which aren't coded the same. This means that
2881 almost everything to do with SETs or SEQUENCEs has either changed name or
2882 number of arguments.
2883 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
2885 *) Fix test data to work with the above.
2888 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
2889 was already fixed by Eric for 0.9.1 it seems.
2890 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
2892 *) Autodetect FreeBSD3.
2895 *) Fix various bugs in Configure. This affects the following platforms:
2899 unixware-2.0-pentium
2903 *) Eliminate generated files from CVS. Reorder tests to regenerate files
2904 before they are needed.
2907 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
2911 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
2913 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
2914 changed SSLeay to OpenSSL in version strings.
2915 [Ralf S. Engelschall]
2917 *) Some fixups to the top-level documents.
2920 *) Fixed the nasty bug where rsaref.h was not found under compile-time
2921 because the symlink to include/ was missing.
2922 [Ralf S. Engelschall]
2924 *) Incorporated the popular no-RSA/DSA-only patches
2925 which allow to compile a RSA-free SSLeay.
2926 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
2928 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
2929 when "ssleay" is still not found.
2930 [Ralf S. Engelschall]
2932 *) Added more platforms to Configure: Cray T3E, HPUX 11,
2933 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
2935 *) Updated the README file.
2936 [Ralf S. Engelschall]
2938 *) Added various .cvsignore files in the CVS repository subdirs
2939 to make a "cvs update" really silent.
2940 [Ralf S. Engelschall]
2942 *) Recompiled the error-definition header files and added
2943 missing symbols to the Win32 linker tables.
2944 [Ralf S. Engelschall]
2946 *) Cleaned up the top-level documents;
2947 o new files: CHANGES and LICENSE
2948 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
2949 o merged COPYRIGHT into LICENSE
2950 o removed obsolete TODO file
2951 o renamed MICROSOFT to INSTALL.W32
2952 [Ralf S. Engelschall]
2954 *) Removed dummy files from the 0.9.1b source tree:
2955 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
2956 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
2957 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
2958 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
2959 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
2960 [Ralf S. Engelschall]
2962 *) Added various platform portability fixes.
2965 *) The Genesis of the OpenSSL rpject:
2966 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
2967 Young and Tim J. Hudson created while they were working for C2Net until
2969 [The OpenSSL Project]
2972 Changes between 0.9.0b and 0.9.1b [not released]
2974 *) Updated a few CA certificates under certs/
2977 *) Changed some BIGNUM api stuff.
2980 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
2981 DGUX x86, Linux Alpha, etc.
2984 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
2985 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
2989 *) Add -strparse option to asn1pars program which parses nested
2991 [Dr Stephen Henson <shenson@bigfoot.com>]
2993 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
2996 *) DSA fix for "ca" program.
2999 *) Added "-genkey" option to "dsaparam" program.
3002 *) Added RIPE MD160 (rmd160) message digest.
3005 *) Added -a (all) option to "ssleay version" command.
3008 *) Added PLATFORM define which is the id given to Configure.
3011 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
3014 *) Extended the ASN.1 parser routines.
3017 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
3020 *) Added a BN_CTX to the BN library.
3023 *) Fixed the weak key values in DES library
3026 *) Changed API in EVP library for cipher aliases.
3029 *) Added support for RC2/64bit cipher.
3032 *) Converted the lhash library to the crypto/mem.c functions.
3035 *) Added more recognized ASN.1 object ids.
3038 *) Added more RSA padding checks for SSL/TLS.
3041 *) Added BIO proxy/filter functionality.
3044 *) Added extra_certs to SSL_CTX which can be used
3045 send extra CA certificates to the client in the CA cert chain sending
3046 process. It can be configured with SSL_CTX_add_extra_chain_cert().
3049 *) Now Fortezza is denied in the authentication phase because
3050 this is key exchange mechanism is not supported by SSLeay at all.
3053 *) Additional PKCS1 checks.
3056 *) Support the string "TLSv1" for all TLS v1 ciphers.
3059 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
3060 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
3063 *) Fixed a few memory leaks.
3066 *) Fixed various code and comment typos.
3069 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
3070 bytes sent in the client random.
3071 [Edward Bishop <ebishop@spyglass.com>]