5 Changes between 0.9.6 and 0.9.6a [xx XXX 2001]
7 *) Fix PKCS#7 decode routines so they correctly update the length
8 after reading an EOC for the EXPLICIT tag.
10 [This change does not apply to 0.9.7.]
12 *) Fix bug in PKCS#12 key generation routines. This was triggered
13 if a 3DES key was generated with a 0 initial byte. Include
14 PKCS12_BROKEN_KEYGEN compilation option to retain the old
15 (but broken) behaviour.
18 *) Fix memory leaks in err.c: free err_data string if necessary;
19 don't write to the wrong index in ERR_set_error_data.
22 *) Implement ssl23_peek (analogous to ssl23_read), which previously
26 *) Replace rdtsc with _emit statements for VC++ version 5.
27 [Jeremy Cooper <jeremy@baymoo.org>]
29 *) Make it possible to reuse SSLv2 sessions.
32 *) In copy_email() check for >= 0 as a return value for
33 X509_NAME_get_index_by_NID() since 0 is a valid index.
34 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
36 *) Avoid coredump with unsupported or invalid public keys by checking if
37 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
38 PKCS7_verify() fails with non detached data.
41 *) Don't use getenv in library functions when run as setuid/setgid.
42 New function OPENSSL_issetugid().
45 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
46 due to incorrect handling of multi-threading:
48 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
50 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
52 3. Count how many times MemCheck_off() has been called so that
53 nested use can be treated correctly. This also avoids
54 inband-signalling in the previous code (which relied on the
55 assumption that thread ID 0 is impossible).
58 *) Add "-rand" option also to s_client and s_server.
61 *) Fix CPU detection on Irix 6.x.
62 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
63 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
65 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
68 [This change does not apply to 0.9.7.]
70 *) Use the cached encoding of an X509_NAME structure rather than
71 copying it. This is apparently the reason for the libsafe "errors"
72 but the code is actually correct.
75 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
76 Bleichenbacher's DSA attack.
77 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
78 to be set and top=0 forces the highest bit to be set; top=-1 is new
79 and leaves the highest bit random.
80 [Ulf Moeller, Bodo Moeller]
82 *) In the NCONF_...-based implementations for CONF_... queries
83 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
84 a temporary CONF structure with the data component set to NULL
85 (which gives segmentation faults in lh_retrieve).
86 Instead, use NULL for the CONF pointer in CONF_get_string and
87 CONF_get_number (which may use environment variables) and directly
88 return NULL from CONF_get_section.
91 *) Fix potential buffer overrun for EBCDIC.
94 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
95 keyUsage if basicConstraints absent for a CA.
98 *) Make SMIME_write_PKCS7() write mail header values with a format that
99 is more generally accepted (no spaces before the semicolon), since
100 some programs can't parse those values properly otherwise. Also make
101 sure BIO's that break lines after each write do not create invalid
105 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
106 macros previously used would not encode an empty SEQUENCE OF
107 and break the signature.
109 [This change does not apply to 0.9.7.]
111 *) Zero the premaster secret after deriving the master secret in
115 *) Add some EVP_add_digest_alias registrations (as found in
116 OpenSSL_add_all_digests()) to SSL_library_init()
117 aka OpenSSL_add_ssl_algorithms(). This provides improved
118 compatibility with peers using X.509 certificates
119 with unconventional AlgorithmIdentifier OIDs.
122 *) Fix for Irix with NO_ASM.
123 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
125 *) ./config script fixes.
126 [Ulf Moeller, Richard Levitte]
128 *) Fix 'openssl passwd -1'.
131 *) Change PKCS12_key_gen_asc() so it can cope with non null
132 terminated strings whose length is passed in the passlen
133 parameter, for example from PEM callbacks. This was done
134 by adding an extra length parameter to asc2uni().
135 [Steve Henson, reported by <oddissey@samsung.co.kr>]
137 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
138 call failed, free the DSA structure.
141 *) Fix to uni2asc() to cope with zero length Unicode strings.
142 These are present in some PKCS#12 files.
145 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
146 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
147 when writing a 32767 byte record.
148 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
150 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
151 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
153 (RSA objects have a reference count access to which is protected
154 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
155 so they are meant to be shared between threads.)
156 [Bodo Moeller, Geoff Thorpe; original patch submitted by
157 "Reddie, Steven" <Steven.Reddie@ca.com>]
159 *) Fix a deadlock in CRYPTO_mem_leaks().
162 *) Use better test patterns in bntest.
165 *) rand_win.c fix for Borland C.
168 *) BN_rshift bugfix for n == 0.
171 *) Add a 'bctest' script that checks for some known 'bc' bugs
172 so that 'make test' does not abort just because 'bc' is broken.
175 *) Store verify_result within SSL_SESSION also for client side to
176 avoid potential security hole. (Re-used sessions on the client side
177 always resulted in verify_result==X509_V_OK, not using the original
178 result of the server certificate verification.)
181 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
182 SSL3_RT_APPLICATION_DATA, return 0.
183 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
187 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
188 releases, have been re-implemented by renaming the previous
189 implementations of ssl2_read and ssl3_read to ssl2_read_internal
190 and ssl3_read_internal, respectively, and adding 'peek' parameters
191 to them. The new ssl[23]_{read,peek} functions are calls to
192 ssl[23]_read_internal with the 'peek' flag set appropriately.
193 A 'peek' parameter has also been added to ssl3_read_bytes, which
194 does the actual work for ssl3_read_internal.
197 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
198 the method-specific "init()" handler. Also clean up ex_data after
199 calling the method-specific "finish()" handler. Previously, this was
200 happening the other way round.
203 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
204 The previous value, 12, was not always sufficient for BN_mod_exp().
207 *) Fix typo in get_cert_by_subject() in by_dir.c
208 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
210 *) Rework the system to generate shared libraries:
212 - Make note of the expected extension for the shared libraries and
213 if there is a need for symbolic links from for example libcrypto.so.0
214 to libcrypto.so.0.9.7. There is extended info in Configure for
217 - Make as few rebuilds of the shared libraries as possible.
219 - Still avoid linking the OpenSSL programs with the shared libraries.
221 - When installing, install the shared libraries separately from the
225 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
227 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
228 and not in SSL_clear because the latter is also used by the
229 accept/connect functions; previously, the settings made by
230 SSL_set_read_ahead would be lost during the handshake.
231 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
233 *) Correct util/mkdef.pl to be selective about disabled algorithms.
234 Previously, it would create entries for disableed algorithms no
238 *) Added several new manual pages for SSL_* function.
241 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
243 *) In ssl23_get_client_hello, generate an error message when faced
244 with an initial SSL 3.0/TLS record that is too small to contain the
245 first two bytes of the ClientHello message, i.e. client_version.
246 (Note that this is a pathologic case that probably has never happened
247 in real life.) The previous approach was to use the version number
248 from the record header as a substitute; but our protocol choice
249 should not depend on that one because it is not authenticated
250 by the Finished messages.
253 *) More robust randomness gathering functions for Windows.
254 [Jeffrey Altman <jaltman@columbia.edu>]
256 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
257 not set then we don't setup the error code for issuer check errors
258 to avoid possibly overwriting other errors which the callback does
259 handle. If an application does set the flag then we assume it knows
260 what it is doing and can handle the new informational codes
264 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
265 a general "ANY" type, as such it should be able to decode anything
266 including tagged types. However it didn't check the class so it would
267 wrongly interpret tagged types in the same way as their universal
268 counterpart and unknown types were just rejected. Changed so that the
269 tagged and unknown types are handled in the same way as a SEQUENCE:
270 that is the encoding is stored intact. There is also a new type
271 "V_ASN1_OTHER" which is used when the class is not universal, in this
272 case we have no idea what the actual type is so we just lump them all
276 *) On VMS, stdout may very well lead to a file that is written to
277 in a record-oriented fashion. That means that every write() will
278 write a separate record, which will be read separately by the
279 programs trying to read from it. This can be very confusing.
281 The solution is to put a BIO filter in the way that will buffer
282 text until a linefeed is reached, and then write everything a
283 line at a time, so every record written will be an actual line,
284 not chunks of lines and not (usually doesn't happen, but I've
285 seen it once) several lines in one record. BIO_f_linebuffer() is
288 Currently, it's a VMS-only method, because that's where it has
289 been tested well enough.
292 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
293 it can return incorrect results.
294 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
295 but it was in 0.9.6-beta[12].)
298 *) Disable the check for content being present when verifying detached
299 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
300 include zero length content when signing messages.
303 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
304 BIO_ctrl (for BIO pairs).
307 *) Add DSO method for VMS.
310 *) Bug fix: Montgomery multiplication could produce results with the
314 *) Add RPM specification openssl.spec and modify it to build three
315 packages. The default package contains applications, application
316 documentation and run-time libraries. The devel package contains
317 include files, static libraries and function documentation. The
318 doc package contains the contents of the doc directory. The original
319 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
322 *) Add a large number of documentation files for many SSL routines.
323 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
325 *) Add a configuration entry for Sony News 4.
326 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
328 *) Don't set the two most significant bits to one when generating a
329 random number < q in the DSA library.
332 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
333 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
334 the underlying transport is blocking) if a handshake took place.
335 (The default behaviour is needed by applications such as s_client
336 and s_server that use select() to determine when to use SSL_read;
337 but for applications that know in advance when to expect data, it
338 just makes things more complicated.)
341 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
345 *) Add a few more EBCDIC conditionals that make `req' and `x509'
346 work better on such systems.
347 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
349 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
350 Update PKCS12_parse() so it copies the friendlyName and the
351 keyid to the certificates aux info.
354 *) Fix bug in PKCS7_verify() which caused an infinite loop
355 if there was more than one signature.
356 [Sven Uszpelkat <su@celocom.de>]
358 *) Major change in util/mkdef.pl to include extra information
359 about each symbol, as well as presentig variables as well
360 as functions. This change means that there's n more need
361 to rebuild the .num files when some algorithms are excluded.
364 *) Allow the verify time to be set by an application,
365 rather than always using the current time.
368 *) Phase 2 verify code reorganisation. The certificate
369 verify code now looks up an issuer certificate by a
370 number of criteria: subject name, authority key id
371 and key usage. It also verifies self signed certificates
372 by the same criteria. The main comparison function is
373 X509_check_issued() which performs these checks.
375 Lot of changes were necessary in order to support this
376 without completely rewriting the lookup code.
378 Authority and subject key identifier are now cached.
380 The LHASH 'certs' is X509_STORE has now been replaced
381 by a STACK_OF(X509_OBJECT). This is mainly because an
382 LHASH can't store or retrieve multiple objects with
385 As a result various functions (which were all internal
386 use only) have changed to handle the new X509_STORE
387 structure. This will break anything that messed round
388 with X509_STORE internally.
390 The functions X509_STORE_add_cert() now checks for an
391 exact match, rather than just subject name.
393 The X509_STORE API doesn't directly support the retrieval
394 of multiple certificates matching a given criteria, however
395 this can be worked round by performing a lookup first
396 (which will fill the cache with candidate certificates)
397 and then examining the cache for matches. This is probably
398 the best we can do without throwing out X509_LOOKUP
399 entirely (maybe later...).
401 The X509_VERIFY_CTX structure has been enhanced considerably.
403 All certificate lookup operations now go via a get_issuer()
404 callback. Although this currently uses an X509_STORE it
405 can be replaced by custom lookups. This is a simple way
406 to bypass the X509_STORE hackery necessary to make this
407 work and makes it possible to use more efficient techniques
408 in future. A very simple version which uses a simple
409 STACK for its trusted certificate store is also provided
410 using X509_STORE_CTX_trusted_stack().
412 The verify_cb() and verify() callbacks now have equivalents
413 in the X509_STORE_CTX structure.
415 X509_STORE_CTX also has a 'flags' field which can be used
416 to customise the verify behaviour.
419 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
420 excludes S/MIME capabilities.
423 *) When a certificate request is read in keep a copy of the
424 original encoding of the signed data and use it when outputing
425 again. Signatures then use the original encoding rather than
426 a decoded, encoded version which may cause problems if the
427 request is improperly encoded.
430 *) For consistency with other BIO_puts implementations, call
431 buffer_write(b, ...) directly in buffer_puts instead of calling
434 In BIO_puts, increment b->num_write as in BIO_write.
435 [Peter.Sylvester@EdelWeb.fr]
437 *) Fix BN_mul_word for the case where the word is 0. (We have to use
438 BN_zero, we may not return a BIGNUM with an array consisting of
442 *) Avoid calling abort() from within the library when problems are
443 detected, except if preprocessor symbols have been defined
444 (such as REF_CHECK, BN_DEBUG etc.).
447 *) New openssl application 'rsautl'. This utility can be
448 used for low level RSA operations. DER public key
449 BIO/fp routines also added.
452 *) New Configure entry and patches for compiling on QNX 4.
453 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
455 *) A demo state-machine implementation was sponsored by
456 Nuron (http://www.nuron.com/) and is now available in
460 *) New options added to the 'dgst' utility for signature
461 generation and verification.
464 *) Unrecognized PKCS#7 content types are now handled via a
465 catch all ASN1_TYPE structure. This allows unsupported
466 types to be stored as a "blob" and an application can
467 encode and decode it manually.
470 *) Fix various signed/unsigned issues to make a_strex.c
472 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
474 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
475 length if passed a buffer. ASN1_INTEGER_to_BN failed
476 if passed a NULL BN and its argument was negative.
477 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
479 *) Modification to PKCS#7 encoding routines to output definite
480 length encoding. Since currently the whole structures are in
481 memory there's not real point in using indefinite length
482 constructed encoding. However if OpenSSL is compiled with
483 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
486 *) Added BIO_vprintf() and BIO_vsnprintf().
489 *) Added more prefixes to parse for in the the strings written
490 through a logging bio, to cover all the levels that are available
491 through syslog. The prefixes are now:
493 PANIC, EMERG, EMR => LOG_EMERG
494 ALERT, ALR => LOG_ALERT
495 CRIT, CRI => LOG_CRIT
496 ERROR, ERR => LOG_ERR
497 WARNING, WARN, WAR => LOG_WARNING
498 NOTICE, NOTE, NOT => LOG_NOTICE
499 INFO, INF => LOG_INFO
500 DEBUG, DBG => LOG_DEBUG
502 and as before, if none of those prefixes are present at the
503 beginning of the string, LOG_ERR is chosen.
505 On Win32, the LOG_* levels are mapped according to this:
507 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
508 LOG_WARNING => EVENTLOG_WARNING_TYPE
509 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
513 *) Made it possible to reconfigure with just the configuration
514 argument "reconf" or "reconfigure". The command line arguments
515 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
516 and are retrieved from there when reconfiguring.
520 [Assar Westerlund <assar@sics.se>, Richard Levitte]
522 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
525 *) The obj_dat.pl script was messing up the sorting of object
526 names. The reason was that it compared the quoted version
527 of strings as a result "OCSP" > "OCSP Signing" because
528 " > SPACE. Changed script to store unquoted versions of
529 names and add quotes on output. It was also omitting some
530 names from the lookup table if they were given a default
531 value (that is if SN is missing it is given the same
532 value as LN and vice versa), these are now added on the
533 grounds that if an object has a name we should be able to
534 look it up. Finally added warning output when duplicate
535 short or long names are found.
538 *) Changes needed for Tandem NSK.
539 [Scott Uroff <scott@xypro.com>]
541 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
542 RSA_padding_check_SSLv23(), special padding was never detected
543 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
544 version rollback attacks was not effective.
546 In s23_clnt.c, don't use special rollback-attack detection padding
547 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
548 client; similarly, in s23_srvr.c, don't do the rollback check if
549 SSL 2.0 is the only protocol enabled in the server.
552 *) Make it possible to get hexdumps of unprintable data with 'openssl
553 asn1parse'. By implication, the functions ASN1_parse_dump() and
554 BIO_dump_indent() are added.
557 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
558 these print out strings and name structures based on various
559 flags including RFC2253 support and proper handling of
560 multibyte characters. Added options to the 'x509' utility
561 to allow the various flags to be set.
564 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
565 Also change the functions X509_cmp_current_time() and
566 X509_gmtime_adj() work with an ASN1_TIME structure,
567 this will enable certificates using GeneralizedTime in validity
571 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
572 negative public key encodings) on by default,
573 NO_NEG_PUBKEY_BUG can be set to disable it.
576 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
577 content octets. An i2c_ASN1_OBJECT is unnecessary because
578 the encoding can be trivially obtained from the structure.
581 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
582 not read locks (CRYPTO_r_[un]lock).
585 *) A first attempt at creating official support for shared
586 libraries through configuration. I've kept it so the
587 default is static libraries only, and the OpenSSL programs
588 are always statically linked for now, but there are
589 preparations for dynamic linking in place.
590 This has been tested on Linux and True64.
593 *) Randomness polling function for Win9x, as described in:
594 Peter Gutmann, Software Generation of Practically Strong
598 *) Fix so PRNG is seeded in req if using an already existing
602 *) New options to smime application. -inform and -outform
603 allow alternative formats for the S/MIME message including
604 PEM and DER. The -content option allows the content to be
605 specified separately. This should allow things like Netscape
606 form signing output easier to verify.
609 *) Fix the ASN1 encoding of tags using the 'long form'.
612 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
613 STRING types. These convert content octets to and from the
614 underlying type. The actual tag and length octets are
615 already assumed to have been read in and checked. These
616 are needed because all other string types have virtually
617 identical handling apart from the tag. By having versions
618 of the ASN1 functions that just operate on content octets
619 IMPLICIT tagging can be handled properly. It also allows
620 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
621 and ASN1_INTEGER are identical apart from the tag.
624 *) Change the handling of OID objects as follows:
626 - New object identifiers are inserted in objects.txt, following
627 the syntax given in objects.README.
628 - objects.pl is used to process obj_mac.num and create a new
630 - obj_dat.pl is used to create a new obj_dat.h, using the data in
633 This is currently kind of a hack, and the perl code in objects.pl
634 isn't very elegant, but it works as I intended. The simplest way
635 to check that it worked correctly is to look in obj_dat.h and
636 check the array nid_objs and make sure the objects haven't moved
637 around (this is important!). Additions are OK, as well as
638 consistent name changes.
641 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
644 *) Addition of the command line parameter '-rand file' to 'openssl req'.
645 The given file adds to whatever has already been seeded into the
646 random pool through the RANDFILE configuration file option or
647 environment variable, or the default random state file.
650 *) mkstack.pl now sorts each macro group into lexical order.
651 Previously the output order depended on the order the files
652 appeared in the directory, resulting in needless rewriting
656 *) Patches to make OpenSSL compile under Win32 again. Mostly
657 work arounds for the VC++ problem that it treats func() as
658 func(void). Also stripped out the parts of mkdef.pl that
659 added extra typesafe functions: these no longer exist.
662 *) Reorganisation of the stack code. The macros are now all
663 collected in safestack.h . Each macro is defined in terms of
664 a "stack macro" of the form SKM_<name>(type, a, b). The
665 DEBUG_SAFESTACK is now handled in terms of function casts,
666 this has the advantage of retaining type safety without the
667 use of additional functions. If DEBUG_SAFESTACK is not defined
668 then the non typesafe macros are used instead. Also modified the
669 mkstack.pl script to handle the new form. Needs testing to see
670 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
671 the default if no major problems. Similar behaviour for ASN1_SET_OF
675 *) When some versions of IIS use the 'NET' form of private key the
676 key derivation algorithm is different. Normally MD5(password) is
677 used as a 128 bit RC4 key. In the modified case
678 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
679 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
680 as the old Netscape_RSA functions except they have an additional
681 'sgckey' parameter which uses the modified algorithm. Also added
682 an -sgckey command line option to the rsa utility. Thanks to
683 Adrian Peck <bertie@ncipher.com> for posting details of the modified
684 algorithm to openssl-dev.
687 *) The evp_local.h macros were using 'c.##kname' which resulted in
688 invalid expansion on some systems (SCO 5.0.5 for example).
689 Corrected to 'c.kname'.
690 [Phillip Porch <root@theporch.com>]
692 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
693 a STACK of email addresses from a certificate or request, these look
694 in the subject name and the subject alternative name extensions and
695 omit any duplicate addresses.
698 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
699 This makes DSA verification about 2 % faster.
702 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
703 (meaning that now 2^5 values will be precomputed, which is only 4 KB
704 plus overhead for 1024 bit moduli).
705 This makes exponentiations about 0.5 % faster for 1024 bit
706 exponents (as measured by "openssl speed rsa2048").
709 *) Rename memory handling macros to avoid conflicts with other
711 Malloc => OPENSSL_malloc
712 Malloc_locked => OPENSSL_malloc_locked
713 Realloc => OPENSSL_realloc
717 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
718 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
722 [John Jarvie <jjarvie@newsguy.com>]
724 *) The type-safe stack code has been rejigged. It is now only compiled
725 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
726 by default all type-specific stack functions are "#define"d back to
727 standard stack functions. This results in more streamlined output
728 but retains the type-safety checking possibilities of the original
732 *) The STACK code has been cleaned up, and certain type declarations
733 that didn't make a lot of sense have been brought in line. This has
734 also involved a cleanup of sorts in safestack.h to more correctly
735 map type-safe stack functions onto their plain stack counterparts.
736 This work has also resulted in a variety of "const"ifications of
737 lots of the code, especially "_cmp" operations which should normally
738 be prototyped with "const" parameters anyway.
741 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
742 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
743 (The PRNG state consists of two parts, the large pool 'state' and 'md',
744 where all of 'md' is used each time the PRNG is used, but 'state'
745 is used only indexed by a cyclic counter. As entropy may not be
746 well distributed from the beginning, 'md' is important as a
747 chaining variable. However, the output function chains only half
748 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
749 all of 'md', and seeding with STATE_SIZE dummy bytes will result
750 in all of 'state' being rewritten, with the new values depending
751 on virtually all of 'md'. This overcomes the 80 bit limitation.)
754 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
755 the handshake is continued after ssl_verify_cert_chain();
756 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
757 can lead to 'unexplainable' connection aborts later.
758 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
760 *) Major EVP API cipher revision.
761 Add hooks for extra EVP features. This allows various cipher
762 parameters to be set in the EVP interface. Support added for variable
763 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
764 setting of RC2 and RC5 parameters.
766 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
769 Remove lots of duplicated code from the EVP library. For example *every*
770 cipher init() function handles the 'iv' in the same way according to the
771 cipher mode. They also all do nothing if the 'key' parameter is NULL and
772 for CFB and OFB modes they zero ctx->num.
774 New functionality allows removal of S/MIME code RC2 hack.
776 Most of the routines have the same form and so can be declared in terms
779 By shifting this to the top level EVP_CipherInit() it can be removed from
780 all individual ciphers. If the cipher wants to handle IVs or keys
781 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
784 Change lots of functions like EVP_EncryptUpdate() to now return a
785 value: although software versions of the algorithms cannot fail
786 any installed hardware versions can.
789 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
790 this option is set, tolerate broken clients that send the negotiated
791 protocol version number instead of the requested protocol version
795 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
796 i.e. non-zero for export ciphersuites, zero otherwise.
797 Previous versions had this flag inverted, inconsistent with
798 rsa_tmp_cb (..._TMP_RSA_CB).
799 [Bodo Moeller; problem reported by Amit Chopra]
801 *) Add missing DSA library text string. Work around for some IIS
802 key files with invalid SEQUENCE encoding.
805 *) Add a document (doc/standards.txt) that list all kinds of standards
806 and so on that are implemented in OpenSSL.
809 *) Enhance c_rehash script. Old version would mishandle certificates
810 with the same subject name hash and wouldn't handle CRLs at all.
811 Added -fingerprint option to crl utility, to support new c_rehash
815 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
818 *) Fix for SSL server purpose checking. Server checking was
819 rejecting certificates which had extended key usage present
820 but no ssl client purpose.
821 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
823 *) Make PKCS#12 code work with no password. The PKCS#12 spec
824 is a little unclear about how a blank password is handled.
825 Since the password in encoded as a BMPString with terminating
826 double NULL a zero length password would end up as just the
827 double NULL. However no password at all is different and is
828 handled differently in the PKCS#12 key generation code. NS
829 treats a blank password as zero length. MSIE treats it as no
830 password on export: but it will try both on import. We now do
831 the same: PKCS12_parse() tries zero length and no password if
832 the password is set to "" or NULL (NULL is now a valid password:
833 it wasn't before) as does the pkcs12 application.
836 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
837 perror when PEM_read_bio_X509_REQ fails, the error message must
838 be obtained from the error queue.
841 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
842 it in ERR_remove_state if appropriate, and change ERR_get_state
843 accordingly to avoid race conditions (this is necessary because
844 thread_hash is no longer constant once set).
847 *) Bugfix for linux-elf makefile.one.
850 *) RSA_get_default_method() will now cause a default
851 RSA_METHOD to be chosen if one doesn't exist already.
852 Previously this was only set during a call to RSA_new()
853 or RSA_new_method(NULL) meaning it was possible for
854 RSA_get_default_method() to return NULL.
857 *) Added native name translation to the existing DSO code
858 that will convert (if the flag to do so is set) filenames
859 that are sufficiently small and have no path information
860 into a canonical native form. Eg. "blah" converted to
861 "libblah.so" or "blah.dll" etc.
864 *) New function ERR_error_string_n(e, buf, len) which is like
865 ERR_error_string(e, buf), but writes at most 'len' bytes
866 including the 0 terminator. For ERR_error_string_n, 'buf'
868 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
870 *) CONF library reworked to become more general. A new CONF
871 configuration file reader "class" is implemented as well as a
872 new functions (NCONF_*, for "New CONF") to handle it. The now
873 old CONF_* functions are still there, but are reimplemented to
874 work in terms of the new functions. Also, a set of functions
875 to handle the internal storage of the configuration data is
876 provided to make it easier to write new configuration file
877 reader "classes" (I can definitely see something reading a
878 configuration file in XML format, for example), called _CONF_*,
879 or "the configuration storage API"...
881 The new configuration file reading functions are:
883 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
884 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
886 NCONF_default, NCONF_WIN32
888 NCONF_dump_fp, NCONF_dump_bio
890 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
891 NCONF_new creates a new CONF object. This works in the same way
892 as other interfaces in OpenSSL, like the BIO interface.
893 NCONF_dump_* dump the internal storage of the configuration file,
894 which is useful for debugging. All other functions take the same
895 arguments as the old CONF_* functions wth the exception of the
896 first that must be a `CONF *' instead of a `LHASH *'.
898 To make it easer to use the new classes with the old CONF_* functions,
899 the function CONF_set_default_method is provided.
902 *) Add '-tls1' option to 'openssl ciphers', which was already
903 mentioned in the documentation but had not been implemented.
904 (This option is not yet really useful because even the additional
905 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
908 *) Initial DSO code added into libcrypto for letting OpenSSL (and
909 OpenSSL-based applications) load shared libraries and bind to
910 them in a portable way.
911 [Geoff Thorpe, with contributions from Richard Levitte]
913 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
915 *) Make sure _lrotl and _lrotr are only used with MSVC.
917 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
918 (the default implementation of RAND_status).
920 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
921 to '-clrext' (= clear extensions), as intended and documented.
922 [Bodo Moeller; inconsistency pointed out by Michael Attili
925 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
926 was larger than the MD block size.
927 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
929 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
930 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
931 using the passed key: if the passed key was a private key the result
932 of X509_print(), for example, would be to print out all the private key
936 *) des_quad_cksum() byte order bug fix.
937 [Ulf Möller, using the problem description in krb4-0.9.7, where
938 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
940 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
942 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
944 *) For easily testing in shell scripts whether some command
945 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
946 returns with exit code 0 iff no command of the given name is available.
947 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
948 the output goes to stdout and nothing is printed to stderr.
949 Additional arguments are always ignored.
951 Since for each cipher there is a command of the same name,
952 the 'no-cipher' compilation switches can be tested this way.
954 ('openssl no-XXX' is not able to detect pseudo-commands such
955 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
958 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
961 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
962 is set; it will be thrown away anyway because each handshake creates
964 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
965 to parameters -- in previous versions (since OpenSSL 0.9.3) the
966 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
967 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
970 *) New s_client option -ign_eof: EOF at stdin is ignored, and
971 'Q' and 'R' lose their special meanings (quit/renegotiate).
972 This is part of what -quiet does; unlike -quiet, -ign_eof
973 does not suppress any output.
976 *) Add compatibility options to the purpose and trust code. The
977 purpose X509_PURPOSE_ANY is "any purpose" which automatically
978 accepts a certificate or CA, this was the previous behaviour,
979 with all the associated security issues.
981 X509_TRUST_COMPAT is the old trust behaviour: only and
982 automatically trust self signed roots in certificate store. A
983 new trust setting X509_TRUST_DEFAULT is used to specify that
984 a purpose has no associated trust setting and it should instead
985 use the value in the default purpose.
988 *) Fix the PKCS#8 DSA private key code so it decodes keys again
989 and fix a memory leak.
992 *) In util/mkerr.pl (which implements 'make errors'), preserve
993 reason strings from the previous version of the .c file, as
994 the default to have only downcase letters (and digits) in
995 automatically generated reasons codes is not always appropriate.
998 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
999 using strerror. Previously, ERR_reason_error_string() returned
1000 library names as reason strings for SYSerr; but SYSerr is a special
1001 case where small numbers are errno values, not library numbers.
1004 *) Add '-dsaparam' option to 'openssl dhparam' application. This
1005 converts DSA parameters into DH parameters. (When creating parameters,
1006 DSA_generate_parameters is used.)
1009 *) Include 'length' (recommended exponent length) in C code generated
1010 by 'openssl dhparam -C'.
1013 *) The second argument to set_label in perlasm was already being used
1014 so couldn't be used as a "file scope" flag. Moved to third argument
1018 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
1019 instead of RAND_bytes for encryption IVs and salts.
1022 *) Include RAND_status() into RAND_METHOD instead of implementing
1023 it only for md_rand.c Otherwise replacing the PRNG by calling
1024 RAND_set_rand_method would be impossible.
1027 *) Don't let DSA_generate_key() enter an infinite loop if the random
1028 number generation fails.
1031 *) New 'rand' application for creating pseudo-random output.
1034 *) Added configuration support for Linux/IA64
1035 [Rolf Haberrecker <rolf@suse.de>]
1037 *) Assembler module support for Mingw32.
1040 *) Shared library support for HPUX (in shlib/).
1041 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
1043 *) Shared library support for Solaris gcc.
1044 [Lutz Behnke <behnke@trustcenter.de>]
1046 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
1048 *) PKCS7_encrypt() was adding text MIME headers twice because they
1049 were added manually and by SMIME_crlf_copy().
1052 *) In bntest.c don't call BN_rand with zero bits argument.
1053 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
1055 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
1056 case was implemented. This caused BN_div_recp() to fail occasionally.
1059 *) Add an optional second argument to the set_label() in the perl
1060 assembly language builder. If this argument exists and is set
1061 to 1 it signals that the assembler should use a symbol whose
1062 scope is the entire file, not just the current function. This
1063 is needed with MASM which uses the format label:: for this scope.
1064 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
1066 *) Change the ASN1 types so they are typedefs by default. Before
1067 almost all types were #define'd to ASN1_STRING which was causing
1068 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
1072 *) Change names of new functions to the new get1/get0 naming
1073 convention: After 'get1', the caller owns a reference count
1074 and has to call ..._free; 'get0' returns a pointer to some
1075 data structure without incrementing reference counters.
1076 (Some of the existing 'get' functions increment a reference
1077 counter, some don't.)
1078 Similarly, 'set1' and 'add1' functions increase reference
1079 counters or duplicate objects.
1082 *) Allow for the possibility of temp RSA key generation failure:
1083 the code used to assume it always worked and crashed on failure.
1086 *) Fix potential buffer overrun problem in BIO_printf().
1087 [Ulf Möller, using public domain code by Patrick Powell; problem
1088 pointed out by David Sacerdote <das33@cornell.edu>]
1090 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
1091 RAND_egd() and RAND_status(). In the command line application,
1092 the EGD socket can be specified like a seed file using RANDFILE
1096 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
1097 Some CAs (e.g. Verisign) distribute certificates in this form.
1100 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
1101 list to exclude them. This means that no special compilation option
1102 is needed to use anonymous DH: it just needs to be included in the
1106 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
1107 EVP_MD_type. The old functionality is available in a new macro called
1108 EVP_MD_md(). Change code that uses it and update docs.
1111 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
1112 where the 'void *' argument is replaced by a function pointer argument.
1113 Previously 'void *' was abused to point to functions, which works on
1114 many platforms, but is not correct. As these functions are usually
1115 called by macros defined in OpenSSL header files, most source code
1116 should work without changes.
1119 *) <openssl/opensslconf.h> (which is created by Configure) now contains
1120 sections with information on -D... compiler switches used for
1121 compiling the library so that applications can see them. To enable
1122 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
1123 must be defined. E.g.,
1124 #define OPENSSL_ALGORITHM_DEFINES
1125 #include <openssl/opensslconf.h>
1126 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
1127 [Richard Levitte, Ulf and Bodo Möller]
1129 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
1133 *) Change the 'other' type in certificate aux info to a STACK_OF
1134 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
1135 the required ASN1 format: arbitrary types determined by an OID.
1138 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
1139 argument to 'req'. This is not because the function is newer or
1140 better than others it just uses the work 'NEW' in the certificate
1141 request header lines. Some software needs this.
1144 *) Reorganise password command line arguments: now passwords can be
1145 obtained from various sources. Delete the PEM_cb function and make
1146 it the default behaviour: i.e. if the callback is NULL and the
1147 usrdata argument is not NULL interpret it as a null terminated pass
1148 phrase. If usrdata and the callback are NULL then the pass phrase
1149 is prompted for as usual.
1152 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
1153 the support is automatically enabled. The resulting binaries will
1154 autodetect the card and use it if present.
1155 [Ben Laurie and Compaq Inc.]
1157 *) Work around for Netscape hang bug. This sends certificate request
1158 and server done in one record. Since this is perfectly legal in the
1159 SSL/TLS protocol it isn't a "bug" option and is on by default. See
1160 the bugs/SSLv3 entry for more info.
1163 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
1166 *) Add -rand argument to smime and pkcs12 applications and read/write
1170 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
1173 *) Add command line password options to the remaining applications.
1176 *) Bug fix for BN_div_recp() for numerators with an even number of
1180 *) More tests in bntest.c, and changed test_bn output.
1183 *) ./config recognizes MacOS X now.
1186 *) Bug fix for BN_div() when the first words of num and divsor are
1187 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
1190 *) Add support for various broken PKCS#8 formats, and command line
1191 options to produce them.
1194 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
1195 get temporary BIGNUMs from a BN_CTX.
1198 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
1202 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
1203 include a #define from the old name to the new. The original intent
1204 was that statically linked binaries could for example just call
1205 SSLeay_add_all_ciphers() to just add ciphers to the table and not
1206 link with digests. This never worked becayse SSLeay_add_all_digests()
1207 and SSLeay_add_all_ciphers() were in the same source file so calling
1208 one would link with the other. They are now in separate source files.
1211 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
1214 *) Use a less unusual form of the Miller-Rabin primality test (it used
1215 a binary algorithm for exponentiation integrated into the Miller-Rabin
1216 loop, our standard modexp algorithms are faster).
1219 *) Support for the EBCDIC character set completed.
1220 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
1222 *) Source code cleanups: use const where appropriate, eliminate casts,
1223 use void * instead of char * in lhash.
1226 *) Bugfix: ssl3_send_server_key_exchange was not restartable
1227 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
1228 this the server could overwrite ephemeral keys that the client
1232 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
1233 using 50 iterations of the Rabin-Miller test.
1235 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
1236 iterations of the Rabin-Miller test as required by the appendix
1237 to FIPS PUB 186[-1]) instead of DSA_is_prime.
1238 As BN_is_prime_fasttest includes trial division, DSA parameter
1239 generation becomes much faster.
1241 This implies a change for the callback functions in DSA_is_prime
1242 and DSA_generate_parameters: The callback function is called once
1243 for each positive witness in the Rabin-Miller test, not just
1244 occasionally in the inner loop; and the parameters to the
1245 callback function now provide an iteration count for the outer
1246 loop rather than for the current invocation of the inner loop.
1247 DSA_generate_parameters additionally can call the callback
1248 function with an 'iteration count' of -1, meaning that a
1249 candidate has passed the trial division test (when q is generated
1250 from an application-provided seed, trial division is skipped).
1253 *) New function BN_is_prime_fasttest that optionally does trial
1254 division before starting the Rabin-Miller test and has
1255 an additional BN_CTX * argument (whereas BN_is_prime always
1256 has to allocate at least one BN_CTX).
1257 'callback(1, -1, cb_arg)' is called when a number has passed the
1258 trial division stage.
1261 *) Fix for bug in CRL encoding. The validity dates weren't being handled
1265 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
1268 *) New function BN_pseudo_rand().
1271 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
1272 bignum version of BN_from_montgomery() with the working code from
1273 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
1277 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
1278 made it impossible to use the same SSL_SESSION data structure in
1279 SSL2 clients in multiple threads.
1282 *) The return value of RAND_load_file() no longer counts bytes obtained
1283 by stat(). RAND_load_file(..., -1) is new and uses the complete file
1284 to seed the PRNG (previously an explicit byte count was required).
1285 [Ulf Möller, Bodo Möller]
1287 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
1288 used (char *) instead of (void *) and had casts all over the place.
1291 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
1294 *) Retain source code compatibility for BN_prime_checks macro:
1295 BN_is_prime(..., BN_prime_checks, ...) now uses
1296 BN_prime_checks_for_size to determine the appropriate number of
1297 Rabin-Miller iterations.
1300 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
1301 DH_CHECK_P_NOT_SAFE_PRIME.
1302 (Check if this is true? OpenPGP calls them "strong".)
1305 *) Merge the functionality of "dh" and "gendh" programs into a new program
1306 "dhparam". The old programs are retained for now but will handle DH keys
1307 (instead of parameters) in future.
1310 *) Make the ciphers, s_server and s_client programs check the return values
1311 when a new cipher list is set.
1314 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
1315 ciphers. Before when the 56bit ciphers were enabled the sorting was
1318 The syntax for the cipher sorting has been extended to support sorting by
1319 cipher-strength (using the strength_bits hard coded in the tables).
1320 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
1322 Fix a bug in the cipher-command parser: when supplying a cipher command
1323 string with an "undefined" symbol (neither command nor alphanumeric
1324 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
1325 an error is flagged.
1327 Due to the strength-sorting extension, the code of the
1328 ssl_create_cipher_list() function was completely rearranged. I hope that
1329 the readability was also increased :-)
1330 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
1332 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
1333 for the first serial number and places 2 in the serial number file. This
1334 avoids problems when the root CA is created with serial number zero and
1335 the first user certificate has the same issuer name and serial number
1339 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
1340 the new code. Add documentation for this stuff.
1343 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
1344 X509_*() to X509at_*() on the grounds that they don't handle X509
1345 structures and behave in an analagous way to the X509v3 functions:
1346 they shouldn't be called directly but wrapper functions should be used
1349 So we also now have some wrapper functions that call the X509at functions
1350 when passed certificate requests. (TO DO: similar things can be done with
1351 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
1352 things. Some of these need some d2i or i2d and print functionality
1353 because they handle more complex structures.)
1356 *) Add missing #ifndefs that caused missing symbols when building libssl
1357 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
1358 NO_RSA in ssl/s2*.c.
1359 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
1361 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
1362 has a return value which indicates the quality of the random data
1363 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
1364 error queue. New function RAND_pseudo_bytes() generates output that is
1365 guaranteed to be unique but not unpredictable. RAND_add is like
1366 RAND_seed, but takes an extra argument for an entropy estimate
1367 (RAND_seed always assumes full entropy).
1370 *) Do more iterations of Rabin-Miller probable prime test (specifically,
1371 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
1372 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
1373 in crypto/bn/bn_prime.c for the complete table). This guarantees a
1374 false-positive rate of at most 2^-80 for random input.
1377 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
1380 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
1381 in the 0.9.5 release), this returns the chain
1382 from an X509_CTX structure with a dup of the stack and all
1383 the X509 reference counts upped: so the stack will exist
1384 after X509_CTX_cleanup() has been called. Modify pkcs12.c
1387 Also make SSL_SESSION_print() print out the verify return
1391 *) Add manpage for the pkcs12 command. Also change the default
1392 behaviour so MAC iteration counts are used unless the new
1393 -nomaciter option is used. This improves file security and
1394 only older versions of MSIE (4.0 for example) need it.
1397 *) Honor the no-xxx Configure options when creating .DEF files.
1400 *) Add PKCS#10 attributes to field table: challengePassword,
1401 unstructuredName and unstructuredAddress. These are taken from
1402 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
1403 international characters are used.
1405 More changes to X509_ATTRIBUTE code: allow the setting of types
1406 based on strings. Remove the 'loc' parameter when adding
1407 attributes because these will be a SET OF encoding which is sorted
1411 *) Initial changes to the 'req' utility to allow request generation
1412 automation. This will allow an application to just generate a template
1413 file containing all the field values and have req construct the
1416 Initial support for X509_ATTRIBUTE handling. Stacks of these are
1417 used all over the place including certificate requests and PKCS#7
1418 structures. They are currently handled manually where necessary with
1419 some primitive wrappers for PKCS#7. The new functions behave in a
1420 manner analogous to the X509 extension functions: they allow
1421 attributes to be looked up by NID and added.
1423 Later something similar to the X509V3 code would be desirable to
1424 automatically handle the encoding, decoding and printing of the
1425 more complex types. The string types like challengePassword can
1426 be handled by the string table functions.
1428 Also modified the multi byte string table handling. Now there is
1429 a 'global mask' which masks out certain types. The table itself
1430 can use the flag STABLE_NO_MASK to ignore the mask setting: this
1431 is useful when for example there is only one permissible type
1432 (as in countryName) and using the mask might result in no valid
1436 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
1437 SSL_get_peer_finished to allow applications to obtain the latest
1438 Finished messages sent to the peer or expected from the peer,
1439 respectively. (SSL_get_peer_finished is usually the Finished message
1440 actually received from the peer, otherwise the protocol will be aborted.)
1442 As the Finished message are message digests of the complete handshake
1443 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
1444 be used for external authentication procedures when the authentication
1445 provided by SSL/TLS is not desired or is not enough.
1448 *) Enhanced support for Alpha Linux is added. Now ./config checks if
1449 the host supports BWX extension and if Compaq C is present on the
1450 $PATH. Just exploiting of the BWX extension results in 20-30%
1451 performance kick for some algorithms, e.g. DES and RC4 to mention
1452 a couple. Compaq C in turn generates ~20% faster code for MD5 and
1456 *) Add support for MS "fast SGC". This is arguably a violation of the
1457 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
1458 weak crypto and after checking the certificate is SGC a second one
1459 with strong crypto. MS SGC stops the first handshake after receiving
1460 the server certificate message and sends a second client hello. Since
1461 a server will typically do all the time consuming operations before
1462 expecting any further messages from the client (server key exchange
1463 is the most expensive) there is little difference between the two.
1465 To get OpenSSL to support MS SGC we have to permit a second client
1466 hello message after we have sent server done. In addition we have to
1467 reset the MAC if we do get this second client hello.
1470 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
1471 if a DER encoded private key is RSA or DSA traditional format. Changed
1472 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
1473 format DER encoded private key. Newer code should use PKCS#8 format which
1474 has the key type encoded in the ASN1 structure. Added DER private key
1475 support to pkcs8 application.
1478 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
1479 ciphersuites has been selected (as required by the SSL 3/TLS 1
1480 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
1481 is set, we interpret this as a request to violate the specification
1482 (the worst that can happen is a handshake failure, and 'correct'
1483 behaviour would result in a handshake failure anyway).
1486 *) In SSL_CTX_add_session, take into account that there might be multiple
1487 SSL_SESSION structures with the same session ID (e.g. when two threads
1488 concurrently obtain them from an external cache).
1489 The internal cache can handle only one SSL_SESSION with a given ID,
1490 so if there's a conflict, we now throw out the old one to achieve
1494 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
1495 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
1496 some routines that use cipher OIDs: some ciphers do not have OIDs
1497 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
1501 *) Simplify the trust setting structure and code. Now we just have
1502 two sequences of OIDs for trusted and rejected settings. These will
1503 typically have values the same as the extended key usage extension
1504 and any application specific purposes.
1506 The trust checking code now has a default behaviour: it will just
1507 check for an object with the same NID as the passed id. Functions can
1508 be provided to override either the default behaviour or the behaviour
1509 for a given id. SSL client, server and email already have functions
1510 in place for compatibility: they check the NID and also return "trusted"
1511 if the certificate is self signed.
1514 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
1515 traditional format into an EVP_PKEY structure.
1518 *) Add a password callback function PEM_cb() which either prompts for
1519 a password if usr_data is NULL or otherwise assumes it is a null
1520 terminated password. Allow passwords to be passed on command line
1521 environment or config files in a few more utilities.
1524 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
1525 keys. Add some short names for PKCS#8 PBE algorithms and allow them
1526 to be specified on the command line for the pkcs8 and pkcs12 utilities.
1527 Update documentation.
1530 *) Support for ASN1 "NULL" type. This could be handled before by using
1531 ASN1_TYPE but there wasn't any function that would try to read a NULL
1532 and produce an error if it couldn't. For compatibility we also have
1533 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
1534 don't allocate anything because they don't need to.
1537 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
1539 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
1541 *) Rebuild of the memory allocation routines used by OpenSSL code and
1542 possibly others as well. The purpose is to make an interface that
1543 provide hooks so anyone can build a separate set of allocation and
1544 deallocation routines to be used by OpenSSL, for example memory
1545 pool implementations, or something else, which was previously hard
1546 since Malloc(), Realloc() and Free() were defined as macros having
1547 the values malloc, realloc and free, respectively (except for Win32
1548 compilations). The same is provided for memory debugging code.
1549 OpenSSL already comes with functionality to find memory leaks, but
1550 this gives people a chance to debug other memory problems.
1552 With these changes, a new set of functions and macros have appeared:
1554 CRYPTO_set_mem_debug_functions() [F]
1555 CRYPTO_get_mem_debug_functions() [F]
1556 CRYPTO_dbg_set_options() [F]
1557 CRYPTO_dbg_get_options() [F]
1558 CRYPTO_malloc_debug_init() [M]
1560 The memory debug functions are NULL by default, unless the library
1561 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
1562 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
1563 gives the standard debugging functions that come with OpenSSL) or
1564 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
1565 provided by the library user) must be used. When the standard
1566 debugging functions are used, CRYPTO_dbg_set_options can be used to
1567 request additional information:
1568 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
1569 the CRYPTO_MDEBUG_xxx macro when compiling the library.
1571 Also, things like CRYPTO_set_mem_functions will always give the
1572 expected result (the new set of functions is used for allocation
1573 and deallocation) at all times, regardless of platform and compiler
1576 To finish it up, some functions that were never use in any other
1577 way than through macros have a new API and new semantic:
1580 CRYPTO_dbg_realloc()
1583 All macros of value have retained their old syntax.
1584 [Richard Levitte and Bodo Moeller]
1586 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
1587 ordering of SMIMECapabilities wasn't in "strength order" and there
1588 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
1592 *) Some ASN1 types with illegal zero length encoding (INTEGER,
1593 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
1594 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
1596 *) Merge in my S/MIME library for OpenSSL. This provides a simple
1597 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
1598 functionality to handle multipart/signed properly) and a utility
1599 called 'smime' to call all this stuff. This is based on code I
1600 originally wrote for Celo who have kindly allowed it to be
1601 included in OpenSSL.
1604 *) Add variants des_set_key_checked and des_set_key_unchecked of
1605 des_set_key (aka des_key_sched). Global variable des_check_key
1606 decides which of these is called by des_set_key; this way
1607 des_check_key behaves as it always did, but applications and
1608 the library itself, which was buggy for des_check_key == 1,
1609 have a cleaner way to pick the version they need.
1612 *) New function PKCS12_newpass() which changes the password of a
1616 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
1617 dynamic mix. In both cases the ids can be used as an index into the
1618 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
1619 functions so they accept a list of the field values and the
1620 application doesn't need to directly manipulate the X509_TRUST
1624 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
1628 *) Modify the way the V3 extension code looks up extensions. This now
1629 works in a similar way to the object code: we have some "standard"
1630 extensions in a static table which is searched with OBJ_bsearch()
1631 and the application can add dynamic ones if needed. The file
1632 crypto/x509v3/ext_dat.h now has the info: this file needs to be
1633 updated whenever a new extension is added to the core code and kept
1634 in ext_nid order. There is a simple program 'tabtest.c' which checks
1635 this. New extensions are not added too often so this file can readily
1636 be maintained manually.
1638 There are two big advantages in doing things this way. The extensions
1639 can be looked up immediately and no longer need to be "added" using
1640 X509V3_add_standard_extensions(): this function now does nothing.
1641 [Side note: I get *lots* of email saying the extension code doesn't
1642 work because people forget to call this function]
1643 Also no dynamic allocation is done unless new extensions are added:
1644 so if we don't add custom extensions there is no need to call
1645 X509V3_EXT_cleanup().
1648 *) Modify enc utility's salting as follows: make salting the default. Add a
1649 magic header, so unsalted files fail gracefully instead of just decrypting
1650 to garbage. This is because not salting is a big security hole, so people
1651 should be discouraged from doing it.
1654 *) Fixes and enhancements to the 'x509' utility. It allowed a message
1655 digest to be passed on the command line but it only used this
1656 parameter when signing a certificate. Modified so all relevant
1657 operations are affected by the digest parameter including the
1658 -fingerprint and -x509toreq options. Also -x509toreq choked if a
1659 DSA key was used because it didn't fix the digest.
1662 *) Initial certificate chain verify code. Currently tests the untrusted
1663 certificates for consistency with the verify purpose (which is set
1664 when the X509_STORE_CTX structure is set up) and checks the pathlength.
1666 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
1667 this is because it will reject chains with invalid extensions whereas
1668 every previous version of OpenSSL and SSLeay made no checks at all.
1670 Trust code: checks the root CA for the relevant trust settings. Trust
1671 settings have an initial value consistent with the verify purpose: e.g.
1672 if the verify purpose is for SSL client use it expects the CA to be
1673 trusted for SSL client use. However the default value can be changed to
1674 permit custom trust settings: one example of this would be to only trust
1675 certificates from a specific "secure" set of CAs.
1677 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
1678 which should be used for version portability: especially since the
1679 verify structure is likely to change more often now.
1681 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
1682 to set them. If not set then assume SSL clients will verify SSL servers
1685 Two new options to the verify program: -untrusted allows a set of
1686 untrusted certificates to be passed in and -purpose which sets the
1687 intended purpose of the certificate. If a purpose is set then the
1688 new chain verify code is used to check extension consistency.
1691 *) Support for the authority information access extension.
1694 *) Modify RSA and DSA PEM read routines to transparently handle
1695 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
1696 public keys in a format compatible with certificate
1697 SubjectPublicKeyInfo structures. Unfortunately there were already
1698 functions called *_PublicKey_* which used various odd formats so
1699 these are retained for compatibility: however the DSA variants were
1700 never in a public release so they have been deleted. Changed dsa/rsa
1701 utilities to handle the new format: note no releases ever handled public
1702 keys so we should be OK.
1704 The primary motivation for this change is to avoid the same fiasco
1705 that dogs private keys: there are several incompatible private key
1706 formats some of which are standard and some OpenSSL specific and
1707 require various evil hacks to allow partial transparent handling and
1708 even then it doesn't work with DER formats. Given the option anything
1709 other than PKCS#8 should be dumped: but the other formats have to
1710 stay in the name of compatibility.
1712 With public keys and the benefit of hindsight one standard format
1713 is used which works with EVP_PKEY, RSA or DSA structures: though
1714 it clearly returns an error if you try to read the wrong kind of key.
1716 Added a -pubkey option to the 'x509' utility to output the public key.
1717 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
1718 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
1719 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
1720 that do the same as the EVP_PKEY_assign_*() except they up the
1721 reference count of the added key (they don't "swallow" the
1725 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
1726 CRLs would fail if the file contained no certificates or no CRLs:
1727 added a new function to read in both types and return the number
1728 read: this means that if none are read it will be an error. The
1729 DER versions of the certificate and CRL reader would always fail
1730 because it isn't possible to mix certificates and CRLs in DER format
1731 without choking one or the other routine. Changed this to just read
1732 a certificate: this is the best we can do. Also modified the code
1733 in apps/verify.c to take notice of return codes: it was previously
1734 attempting to read in certificates from NULL pointers and ignoring
1735 any errors: this is one reason why the cert and CRL reader seemed
1736 to work. It doesn't check return codes from the default certificate
1737 routines: these may well fail if the certificates aren't installed.
1740 *) Code to support otherName option in GeneralName.
1743 *) First update to verify code. Change the verify utility
1744 so it warns if it is passed a self signed certificate:
1745 for consistency with the normal behaviour. X509_verify
1746 has been modified to it will now verify a self signed
1747 certificate if *exactly* the same certificate appears
1748 in the store: it was previously impossible to trust a
1749 single self signed certificate. This means that:
1750 openssl verify ss.pem
1751 now gives a warning about a self signed certificate but
1752 openssl verify -CAfile ss.pem ss.pem
1756 *) For servers, store verify_result in SSL_SESSION data structure
1757 (and add it to external session representation).
1758 This is needed when client certificate verifications fails,
1759 but an application-provided verification callback (set by
1760 SSL_CTX_set_cert_verify_callback) allows accepting the session
1761 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
1762 but returns 1): When the session is reused, we have to set
1763 ssl->verify_result to the appropriate error code to avoid
1765 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
1767 *) Fix a bug in the new PKCS#7 code: it didn't consider the
1768 case in PKCS7_dataInit() where the signed PKCS7 structure
1769 didn't contain any existing data because it was being created.
1770 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
1772 *) Add a salt to the key derivation routines in enc.c. This
1773 forms the first 8 bytes of the encrypted file. Also add a
1774 -S option to allow a salt to be input on the command line.
1777 *) New function X509_cmp(). Oddly enough there wasn't a function
1778 to compare two certificates. We do this by working out the SHA1
1779 hash and comparing that. X509_cmp() will be needed by the trust
1783 *) SSL_get1_session() is like SSL_get_session(), but increments
1784 the reference count in the SSL_SESSION returned.
1785 [Geoff Thorpe <geoff@eu.c2.net>]
1787 *) Fix for 'req': it was adding a null to request attributes.
1788 Also change the X509_LOOKUP and X509_INFO code to handle
1789 certificate auxiliary information.
1792 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
1796 *) Add the possibility to add extra information to the memory leak
1797 detecting output, to form tracebacks, showing from where each
1798 allocation was originated: CRYPTO_push_info("constant string") adds
1799 the string plus current file name and line number to a per-thread
1800 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
1801 is like calling CYRPTO_pop_info() until the stack is empty.
1802 Also updated memory leak detection code to be multi-thread-safe.
1805 *) Add options -text and -noout to pkcs7 utility and delete the
1806 encryption options which never did anything. Update docs.
1809 *) Add options to some of the utilities to allow the pass phrase
1810 to be included on either the command line (not recommended on
1811 OSes like Unix) or read from the environment. Update the
1812 manpages and fix a few bugs.
1815 *) Add a few manpages for some of the openssl commands.
1818 *) Fix the -revoke option in ca. It was freeing up memory twice,
1819 leaking and not finding already revoked certificates.
1822 *) Extensive changes to support certificate auxiliary information.
1823 This involves the use of X509_CERT_AUX structure and X509_AUX
1824 functions. An X509_AUX function such as PEM_read_X509_AUX()
1825 can still read in a certificate file in the usual way but it
1826 will also read in any additional "auxiliary information". By
1827 doing things this way a fair degree of compatibility can be
1828 retained: existing certificates can have this information added
1829 using the new 'x509' options.
1831 Current auxiliary information includes an "alias" and some trust
1832 settings. The trust settings will ultimately be used in enhanced
1833 certificate chain verification routines: currently a certificate
1834 can only be trusted if it is self signed and then it is trusted
1838 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
1839 The problem was that one of the replacement routines had not been working
1840 since SSLeay releases. For now the offending routine has been replaced
1841 with non-optimised assembler. Even so, this now gives around 95%
1842 performance improvement for 1024 bit RSA signs.
1845 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
1846 handling. Most clients have the effective key size in bits equal to
1847 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
1848 A few however don't do this and instead use the size of the decrypted key
1849 to determine the RC2 key length and the AlgorithmIdentifier to determine
1850 the effective key length. In this case the effective key length can still
1851 be 40 bits but the key length can be 168 bits for example. This is fixed
1852 by manually forcing an RC2 key into the EVP_PKEY structure because the
1853 EVP code can't currently handle unusual RC2 key sizes: it always assumes
1854 the key length and effective key length are equal.
1857 *) Add a bunch of functions that should simplify the creation of
1858 X509_NAME structures. Now you should be able to do:
1859 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
1860 and have it automatically work out the correct field type and fill in
1861 the structures. The more adventurous can try:
1862 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
1863 and it will (hopefully) work out the correct multibyte encoding.
1866 *) Change the 'req' utility to use the new field handling and multibyte
1867 copy routines. Before the DN field creation was handled in an ad hoc
1868 way in req, ca, and x509 which was rather broken and didn't support
1869 BMPStrings or UTF8Strings. Since some software doesn't implement
1870 BMPStrings or UTF8Strings yet, they can be enabled using the config file
1871 using the dirstring_type option. See the new comment in the default
1872 openssl.cnf for more info.
1875 *) Make crypto/rand/md_rand.c more robust:
1876 - Assure unique random numbers after fork().
1877 - Make sure that concurrent threads access the global counter and
1878 md serializably so that we never lose entropy in them
1879 or use exactly the same state in multiple threads.
1880 Access to the large state is not always serializable because
1881 the additional locking could be a performance killer, and
1882 md should be large enough anyway.
1885 *) New file apps/app_rand.c with commonly needed functionality
1886 for handling the random seed file.
1888 Use the random seed file in some applications that previously did not:
1890 dsaparam -genkey (which also ignored its '-rand' option),
1893 x509 (when signing).
1894 Except on systems with /dev/urandom, it is crucial to have a random
1895 seed file at least for key creation, DSA signing, and for DH exchanges;
1896 for RSA signatures we could do without one.
1898 gendh and gendsa (unlike genrsa) used to read only the first byte
1899 of each file listed in the '-rand' option. The function as previously
1900 found in genrsa is now in app_rand.c and is used by all programs
1901 that support '-rand'.
1904 *) In RAND_write_file, use mode 0600 for creating files;
1905 don't just chmod when it may be too late.
1908 *) Report an error from X509_STORE_load_locations
1909 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
1912 *) New function ASN1_mbstring_copy() this copies a string in either
1913 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
1914 into an ASN1_STRING type. A mask of permissible types is passed
1915 and it chooses the "minimal" type to use or an error if not type
1919 *) Add function equivalents to the various macros in asn1.h. The old
1920 macros are retained with an M_ prefix. Code inside the library can
1921 use the M_ macros. External code (including the openssl utility)
1922 should *NOT* in order to be "shared library friendly".
1925 *) Add various functions that can check a certificate's extensions
1926 to see if it usable for various purposes such as SSL client,
1927 server or S/MIME and CAs of these types. This is currently
1928 VERY EXPERIMENTAL but will ultimately be used for certificate chain
1929 verification. Also added a -purpose flag to x509 utility to
1930 print out all the purposes.
1933 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
1937 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
1938 for, obtain and decode and extension and obtain its critical flag.
1939 This allows all the necessary extension code to be handled in a
1940 single function call.
1943 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
1944 platforms. See crypto/rc4/rc4_enc.c for further details.
1947 *) New -noout option to asn1parse. This causes no output to be produced
1948 its main use is when combined with -strparse and -out to extract data
1949 from a file (which may not be in ASN.1 format).
1952 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
1953 when producing the local key id.
1954 [Richard Levitte <levitte@stacken.kth.se>]
1956 *) New option -dhparam in s_server. This allows a DH parameter file to be
1957 stated explicitly. If it is not stated then it tries the first server
1958 certificate file. The previous behaviour hard coded the filename
1962 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
1963 a public key to be input or output. For example:
1964 openssl rsa -in key.pem -pubout -out pubkey.pem
1965 Also added necessary DSA public key functions to handle this.
1968 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
1969 in the message. This was handled by allowing
1970 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
1971 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
1973 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
1974 to the end of the strings whereas this didn't. This would cause problems
1975 if strings read with d2i_ASN1_bytes() were later modified.
1976 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
1978 *) Fix for base64 decode bug. When a base64 bio reads only one line of
1979 data and it contains EOF it will end up returning an error. This is
1980 caused by input 46 bytes long. The cause is due to the way base64
1981 BIOs find the start of base64 encoded data. They do this by trying a
1982 trial decode on each line until they find one that works. When they
1983 do a flag is set and it starts again knowing it can pass all the
1984 data directly through the decoder. Unfortunately it doesn't reset
1985 the context it uses. This means that if EOF is reached an attempt
1986 is made to pass two EOFs through the context and this causes the
1987 resulting error. This can also cause other problems as well. As is
1988 usual with these problems it takes *ages* to find and the fix is
1989 trivial: move one line.
1990 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
1992 *) Ugly workaround to get s_client and s_server working under Windows. The
1993 old code wouldn't work because it needed to select() on sockets and the
1994 tty (for keypresses and to see if data could be written). Win32 only
1995 supports select() on sockets so we select() with a 1s timeout on the
1996 sockets and then see if any characters are waiting to be read, if none
1997 are present then we retry, we also assume we can always write data to
1998 the tty. This isn't nice because the code then blocks until we've
1999 received a complete line of data and it is effectively polling the
2000 keyboard at 1s intervals: however it's quite a bit better than not
2001 working at all :-) A dedicated Windows application might handle this
2002 with an event loop for example.
2005 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
2006 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
2007 will be called when RSA_sign() and RSA_verify() are used. This is useful
2008 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
2009 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
2010 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
2011 This necessitated the support of an extra signature type NID_md5_sha1
2012 for SSL signatures and modifications to the SSL library to use it instead
2013 of calling RSA_public_decrypt() and RSA_private_encrypt().
2016 *) Add new -verify -CAfile and -CApath options to the crl program, these
2017 will lookup a CRL issuers certificate and verify the signature in a
2018 similar way to the verify program. Tidy up the crl program so it
2019 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
2020 less strict. It will now permit CRL extensions even if it is not
2021 a V2 CRL: this will allow it to tolerate some broken CRLs.
2024 *) Initialize all non-automatic variables each time one of the openssl
2025 sub-programs is started (this is necessary as they may be started
2026 multiple times from the "OpenSSL>" prompt).
2027 [Lennart Bang, Bodo Moeller]
2029 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
2030 removing all other RSA functionality (this is what NO_RSA does). This
2031 is so (for example) those in the US can disable those operations covered
2032 by the RSA patent while allowing storage and parsing of RSA keys and RSA
2036 *) Non-copying interface to BIO pairs.
2037 (still largely untested)
2040 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
2041 ASCII string. This was handled independently in various places before.
2044 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
2045 UTF8 strings a character at a time.
2048 *) Use client_version from client hello to select the protocol
2049 (s23_srvr.c) and for RSA client key exchange verification
2050 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
2053 *) Add various utility functions to handle SPKACs, these were previously
2054 handled by poking round in the structure internals. Added new function
2055 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
2056 print, verify and generate SPKACs. Based on an original idea from
2057 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
2060 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
2063 *) Allow the config file extension section to be overwritten on the
2064 command line. Based on an original idea from Massimiliano Pala
2065 <madwolf@comune.modena.it>. The new option is called -extensions
2066 and can be applied to ca, req and x509. Also -reqexts to override
2067 the request extensions in req and -crlexts to override the crl extensions
2071 *) Add new feature to the SPKAC handling in ca. Now you can include
2072 the same field multiple times by preceding it by "XXXX." for example:
2075 this is the same syntax as used in the req config file.
2078 *) Allow certificate extensions to be added to certificate requests. These
2079 are specified in a 'req_extensions' option of the req section of the
2080 config file. They can be printed out with the -text option to req but
2081 are otherwise ignored at present.
2084 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
2085 data read consists of only the final block it would not decrypted because
2086 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
2087 A misplaced 'break' also meant the decrypted final block might not be
2088 copied until the next read.
2091 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
2092 a few extra parameters to the DH structure: these will be useful if
2093 for example we want the value of 'q' or implement X9.42 DH.
2096 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
2097 provides hooks that allow the default DSA functions or functions on a
2098 "per key" basis to be replaced. This allows hardware acceleration and
2099 hardware key storage to be handled without major modification to the
2100 library. Also added low level modexp hooks and CRYPTO_EX structure and
2101 associated functions.
2104 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
2105 as "read only": it can't be written to and the buffer it points to will
2106 not be freed. Reading from a read only BIO is much more efficient than
2107 a normal memory BIO. This was added because there are several times when
2108 an area of memory needs to be read from a BIO. The previous method was
2109 to create a memory BIO and write the data to it, this results in two
2110 copies of the data and an O(n^2) reading algorithm. There is a new
2111 function BIO_new_mem_buf() which creates a read only memory BIO from
2112 an area of memory. Also modified the PKCS#7 routines to use read only
2116 *) Bugfix: ssl23_get_client_hello did not work properly when called in
2117 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
2118 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
2119 but a retry condition occured while trying to read the rest.
2122 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
2123 NID_pkcs7_encrypted by default: this was wrong since this should almost
2124 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
2125 the encrypted data type: this is a more sensible place to put it and it
2126 allows the PKCS#12 code to be tidied up that duplicated this
2130 *) Changed obj_dat.pl script so it takes its input and output files on
2131 the command line. This should avoid shell escape redirection problems
2135 *) Initial support for certificate extension requests, these are included
2136 in things like Xenroll certificate requests. Included functions to allow
2137 extensions to be obtained and added.
2140 *) -crlf option to s_client and s_server for sending newlines as
2141 CRLF (as required by many protocols).
2144 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
2146 *) Install libRSAglue.a when OpenSSL is built with RSAref.
2147 [Ralf S. Engelschall]
2149 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
2150 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
2152 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
2156 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
2157 DH parameters/keys (q is lost during that conversion, but the resulting
2158 DH parameters contain its length).
2160 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
2161 much faster than DH_generate_parameters (which creates parameters
2162 where p = 2*q + 1), and also the smaller q makes DH computations
2163 much more efficient (160-bit exponentiation instead of 1024-bit
2164 exponentiation); so this provides a convenient way to support DHE
2165 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
2166 utter importance to use
2167 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
2169 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
2170 when such DH parameters are used, because otherwise small subgroup
2171 attacks may become possible!
2174 *) Avoid memory leak in i2d_DHparams.
2177 *) Allow the -k option to be used more than once in the enc program:
2178 this allows the same encrypted message to be read by multiple recipients.
2181 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
2182 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
2183 it will always use the numerical form of the OID, even if it has a short
2187 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
2188 method only got called if p,q,dmp1,dmq1,iqmp components were present,
2189 otherwise bn_mod_exp was called. In the case of hardware keys for example
2190 no private key components need be present and it might store extra data
2191 in the RSA structure, which cannot be accessed from bn_mod_exp.
2192 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
2193 private key operations.
2196 *) Added support for SPARC Linux.
2199 *) pem_password_cb function type incompatibly changed from
2200 typedef int pem_password_cb(char *buf, int size, int rwflag);
2202 ....(char *buf, int size, int rwflag, void *userdata);
2203 so that applications can pass data to their callbacks:
2204 The PEM[_ASN1]_{read,write}... functions and macros now take an
2205 additional void * argument, which is just handed through whenever
2206 the password callback is called.
2207 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
2209 New function SSL_CTX_set_default_passwd_cb_userdata.
2211 Compatibility note: As many C implementations push function arguments
2212 onto the stack in reverse order, the new library version is likely to
2213 interoperate with programs that have been compiled with the old
2214 pem_password_cb definition (PEM_whatever takes some data that
2215 happens to be on the stack as its last argument, and the callback
2216 just ignores this garbage); but there is no guarantee whatsoever that
2219 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
2220 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
2221 problems not only on Windows, but also on some Unix platforms.
2222 To avoid problematic command lines, these definitions are now in an
2223 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
2224 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
2227 *) MIPS III/IV assembler module is reimplemented.
2230 *) More DES library cleanups: remove references to srand/rand and
2231 delete an unused file.
2234 *) Add support for the the free Netwide assembler (NASM) under Win32,
2235 since not many people have MASM (ml) and it can be hard to obtain.
2236 This is currently experimental but it seems to work OK and pass all
2237 the tests. Check out INSTALL.W32 for info.
2240 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
2241 without temporary keys kept an extra copy of the server key,
2242 and connections with temporary keys did not free everything in case
2246 *) New function RSA_check_key and new openssl rsa option -check
2247 for verifying the consistency of RSA keys.
2248 [Ulf Moeller, Bodo Moeller]
2250 *) Various changes to make Win32 compile work:
2251 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
2252 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
2253 comparison" warnings.
2254 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
2257 *) Add a debugging option to PKCS#5 v2 key generation function: when
2258 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
2259 derived keys are printed to stderr.
2262 *) Copy the flags in ASN1_STRING_dup().
2263 [Roman E. Pavlov <pre@mo.msk.ru>]
2265 *) The x509 application mishandled signing requests containing DSA
2266 keys when the signing key was also DSA and the parameters didn't match.
2268 It was supposed to omit the parameters when they matched the signing key:
2269 the verifying software was then supposed to automatically use the CA's
2270 parameters if they were absent from the end user certificate.
2272 Omitting parameters is no longer recommended. The test was also
2273 the wrong way round! This was probably due to unusual behaviour in
2274 EVP_cmp_parameters() which returns 1 if the parameters match.
2275 This meant that parameters were omitted when they *didn't* match and
2276 the certificate was useless. Certificates signed with 'ca' didn't have
2278 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
2280 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
2281 The interface is as follows:
2282 Applications can use
2283 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
2284 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
2285 "off" is now the default.
2286 The library internally uses
2287 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
2288 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
2289 to disable memory-checking temporarily.
2291 Some inconsistent states that previously were possible (and were
2292 even the default) are now avoided.
2294 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
2295 with each memory chunk allocated; this is occasionally more helpful
2296 than just having a counter.
2298 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
2300 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
2304 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
2305 which largely parallels "options", but is for changing API behaviour,
2306 whereas "options" are about protocol behaviour.
2307 Initial "mode" flags are:
2309 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
2310 a single record has been written.
2311 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
2312 retries use the same buffer location.
2313 (But all of the contents must be
2317 *) Bugfix: SSL_set_mode ignored its parameter, only SSL_CTX_set_mode
2320 *) Fix problems with no-hmac etc.
2321 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
2323 *) New functions RSA_get_default_method(), RSA_set_method() and
2324 RSA_get_method(). These allows replacement of RSA_METHODs without having
2325 to mess around with the internals of an RSA structure.
2328 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
2329 Also really enable memory leak checks in openssl.c and in some
2331 [Chad C. Mulligan, Bodo Moeller]
2333 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
2334 up the length of negative integers. This has now been simplified to just
2335 store the length when it is first determined and use it later, rather
2336 than trying to keep track of where data is copied and updating it to
2338 [Steve Henson, reported by Brien Wheeler
2339 <bwheeler@authentica-security.com>]
2341 *) Add a new function PKCS7_signatureVerify. This allows the verification
2342 of a PKCS#7 signature but with the signing certificate passed to the
2343 function itself. This contrasts with PKCS7_dataVerify which assumes the
2344 certificate is present in the PKCS#7 structure. This isn't always the
2345 case: certificates can be omitted from a PKCS#7 structure and be
2346 distributed by "out of band" means (such as a certificate database).
2349 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
2350 function prototypes in pem.h, also change util/mkdef.pl to add the
2351 necessary function names.
2354 *) mk1mf.pl (used by Windows builds) did not properly read the
2355 options set by Configure in the top level Makefile, and Configure
2356 was not even able to write more than one option correctly.
2357 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
2360 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
2361 file to be loaded from a BIO or FILE pointer. The BIO version will
2362 for example allow memory BIOs to contain config info.
2365 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
2366 Whoever hopes to achieve shared-library compatibility across versions
2367 must use this, not the compile-time macro.
2368 (Exercise 0.9.4: Which is the minimum library version required by
2370 Note: All this applies only to multi-threaded programs, others don't
2374 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
2375 through a BIO pair triggered the default case, i.e.
2376 SSLerr(...,SSL_R_UNKNOWN_STATE).
2379 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
2380 can use the SSL library even if none of the specific BIOs is
2384 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
2385 for the encoded length.
2386 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
2388 *) Add initial documentation of the X509V3 functions.
2391 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
2392 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
2393 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
2394 secure PKCS#8 private key format with a high iteration count.
2397 *) Fix determination of Perl interpreter: A perl or perl5
2398 _directory_ in $PATH was also accepted as the interpreter.
2399 [Ralf S. Engelschall]
2401 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
2402 wrong with it but it was very old and did things like calling
2403 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
2407 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
2408 to use the new extension code.
2411 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
2412 with macros. This should make it easier to change their form, add extra
2413 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
2417 *) Add to configuration table a new entry that can specify an alternative
2418 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
2419 according to Mark Crispin <MRC@Panda.COM>.
2423 *) DES CBC did not update the IV. Weird.
2426 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
2427 Changing the behaviour of the former might break existing programs --
2428 where IV updating is needed, des_ncbc_encrypt can be used.
2431 *) When bntest is run from "make test" it drives bc to check its
2432 calculations, as well as internally checking them. If an internal check
2433 fails, it needs to cause bc to give a non-zero result or make test carries
2434 on without noticing the failure. Fixed.
2437 *) DES library cleanups.
2440 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
2441 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
2442 ciphers. NOTE: although the key derivation function has been verified
2443 against some published test vectors it has not been extensively tested
2444 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
2448 *) Instead of "mkdir -p", which is not fully portable, use new
2449 Perl script "util/mkdir-p.pl".
2452 *) Rewrite the way password based encryption (PBE) is handled. It used to
2453 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
2454 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
2455 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
2456 the 'parameter' field of the AlgorithmIdentifier is passed to the
2457 underlying key generation function so it must do its own ASN1 parsing.
2458 This has also changed the EVP_PBE_CipherInit() function which now has a
2459 'parameter' argument instead of literal salt and iteration count values
2460 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
2463 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
2464 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
2465 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
2466 KEY" because this clashed with PKCS#8 unencrypted string. Since this
2467 value was just used as a "magic string" and not used directly its
2468 value doesn't matter.
2471 *) Introduce some semblance of const correctness to BN. Shame C doesn't
2475 *) "linux-sparc64" configuration (ultrapenguin).
2476 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
2477 "linux-sparc" configuration.
2478 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
2480 *) config now generates no-xxx options for missing ciphers.
2483 *) Support the EBCDIC character set (work in progress).
2484 File ebcdic.c not yet included because it has a different license.
2485 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
2487 *) Support BS2000/OSD-POSIX.
2488 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
2490 *) Make callbacks for key generation use void * instead of char *.
2493 *) Make S/MIME samples compile (not yet tested).
2496 *) Additional typesafe stacks.
2499 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
2503 Changes between 0.9.3 and 0.9.3a [29 May 1999]
2505 *) New configuration variant "sco5-gcc".
2507 *) Updated some demos.
2508 [Sean O Riordain, Wade Scholine]
2510 *) Add missing BIO_free at exit of pkcs12 application.
2513 *) Fix memory leak in conf.c.
2516 *) Updates for Win32 to assembler version of MD5.
2519 *) Set #! path to perl in apps/der_chop to where we found it
2520 instead of using a fixed path.
2523 *) SHA library changes for irix64-mips4-cc.
2526 *) Improvements for VMS support.
2530 Changes between 0.9.2b and 0.9.3 [24 May 1999]
2532 *) Bignum library bug fix. IRIX 6 passes "make test" now!
2533 This also avoids the problems with SC4.2 and unpatched SC5.
2534 [Andy Polyakov <appro@fy.chalmers.se>]
2536 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
2537 These are required because of the typesafe stack would otherwise break
2538 existing code. If old code used a structure member which used to be STACK
2539 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
2540 sk_num or sk_value it would produce an error because the num, data members
2541 are not present in STACK_OF. Now it just produces a warning. sk_set
2542 replaces the old method of assigning a value to sk_value
2543 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
2544 that does this will no longer work (and should use sk_set instead) but
2545 this could be regarded as a "questionable" behaviour anyway.
2548 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
2549 correctly handle encrypted S/MIME data.
2552 *) Change type of various DES function arguments from des_cblock
2553 (which means, in function argument declarations, pointer to char)
2554 to des_cblock * (meaning pointer to array with 8 char elements),
2555 which allows the compiler to do more typechecking; it was like
2556 that back in SSLeay, but with lots of ugly casts.
2558 Introduce new type const_des_cblock.
2561 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
2562 problems: find RecipientInfo structure that matches recipient certificate
2563 and initialise the ASN1 structures properly based on passed cipher.
2566 *) Belatedly make the BN tests actually check the results.
2569 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
2570 to and from BNs: it was completely broken. New compilation option
2571 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
2572 key elements as negative integers.
2575 *) Reorganize and speed up MD5.
2576 [Andy Polyakov <appro@fy.chalmers.se>]
2579 [Richard Levitte <richard@levitte.org>]
2581 *) New option -out to asn1parse to allow the parsed structure to be
2582 output to a file. This is most useful when combined with the -strparse
2583 option to examine the output of things like OCTET STRINGS.
2586 *) Make SSL library a little more fool-proof by not requiring any longer
2587 that SSL_set_{accept,connect}_state be called before
2588 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
2589 in many applications because usually everything *appeared* to work as
2590 intended anyway -- now it really works as intended).
2593 *) Move openssl.cnf out of lib/.
2596 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
2597 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
2598 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
2599 [Ralf S. Engelschall]
2601 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
2602 handle PKCS#7 enveloped data properly.
2603 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
2605 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
2606 copying pointers. The cert_st handling is changed by this in
2607 various ways (and thus what used to be known as ctx->default_cert
2608 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
2609 any longer when s->cert does not give us what we need).
2610 ssl_cert_instantiate becomes obsolete by this change.
2611 As soon as we've got the new code right (possibly it already is?),
2612 we have solved a couple of bugs of the earlier code where s->cert
2613 was used as if it could not have been shared with other SSL structures.
2615 Note that using the SSL API in certain dirty ways now will result
2616 in different behaviour than observed with earlier library versions:
2617 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
2618 does not influence s as it used to.
2620 In order to clean up things more thoroughly, inside SSL_SESSION
2621 we don't use CERT any longer, but a new structure SESS_CERT
2622 that holds per-session data (if available); currently, this is
2623 the peer's certificate chain and, for clients, the server's certificate
2624 and temporary key. CERT holds only those values that can have
2625 meaningful defaults in an SSL_CTX.
2628 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
2629 from the internal representation. Various PKCS#7 fixes: remove some
2630 evil casts and set the enc_dig_alg field properly based on the signing
2634 *) Allow PKCS#12 password to be set from the command line or the
2635 environment. Let 'ca' get its config file name from the environment
2636 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
2640 *) Allow certificate policies extension to use an IA5STRING for the
2641 organization field. This is contrary to the PKIX definition but
2642 VeriSign uses it and IE5 only recognises this form. Document 'x509'
2646 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
2647 without disallowing inline assembler and the like for non-pedantic builds.
2650 *) Support Borland C++ builder.
2651 [Janez Jere <jj@void.si>, modified by Ulf Möller]
2656 *) SHA-1 cleanups and performance enhancements.
2657 [Andy Polyakov <appro@fy.chalmers.se>]
2659 *) Sparc v8plus assembler for the bignum library.
2660 [Andy Polyakov <appro@fy.chalmers.se>]
2662 *) Accept any -xxx and +xxx compiler options in Configure.
2665 *) Update HPUX configuration.
2668 *) Add missing sk_<type>_unshift() function to safestack.h
2669 [Ralf S. Engelschall]
2671 *) New function SSL_CTX_use_certificate_chain_file that sets the
2672 "extra_cert"s in addition to the certificate. (This makes sense
2673 only for "PEM" format files, as chains as a whole are not
2677 *) Support verify_depth from the SSL API.
2678 x509_vfy.c had what can be considered an off-by-one-error:
2679 Its depth (which was not part of the external interface)
2680 was actually counting the number of certificates in a chain;
2681 now it really counts the depth.
2684 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
2685 instead of X509err, which often resulted in confusing error
2686 messages since the error codes are not globally unique
2687 (e.g. an alleged error in ssl3_accept when a certificate
2688 didn't match the private key).
2690 *) New function SSL_CTX_set_session_id_context that allows to set a default
2691 value (so that you don't need SSL_set_session_id_context for each
2692 connection using the SSL_CTX).
2695 *) OAEP decoding bug fix.
2698 *) Support INSTALL_PREFIX for package builders, as proposed by
2702 *) New Configure options "threads" and "no-threads". For systems
2703 where the proper compiler options are known (currently Solaris
2704 and Linux), "threads" is the default.
2707 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
2710 *) Install various scripts to $(OPENSSLDIR)/misc, not to
2711 $(INSTALLTOP)/bin -- they shouldn't clutter directories
2712 such as /usr/local/bin.
2715 *) "make linux-shared" to build shared libraries.
2716 [Niels Poppe <niels@netbox.org>]
2718 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
2721 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
2722 extension adding in x509 utility.
2725 *) Remove NOPROTO sections and error code comments.
2728 *) Partial rewrite of the DEF file generator to now parse the ANSI
2732 *) New Configure options --prefix=DIR and --openssldir=DIR.
2735 *) Complete rewrite of the error code script(s). It is all now handled
2736 by one script at the top level which handles error code gathering,
2737 header rewriting and C source file generation. It should be much better
2738 than the old method: it now uses a modified version of Ulf's parser to
2739 read the ANSI prototypes in all header files (thus the old K&R definitions
2740 aren't needed for error creation any more) and do a better job of
2741 translating function codes into names. The old 'ASN1 error code imbedded
2742 in a comment' is no longer necessary and it doesn't use .err files which
2743 have now been deleted. Also the error code call doesn't have to appear all
2744 on one line (which resulted in some large lines...).
2747 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
2750 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
2751 0 (which usually indicates a closed connection), but continue reading.
2754 *) Fix some race conditions.
2757 *) Add support for CRL distribution points extension. Add Certificate
2758 Policies and CRL distribution points documentation.
2761 *) Move the autogenerated header file parts to crypto/opensslconf.h.
2764 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
2765 8 of keying material. Merlin has also confirmed interop with this fix
2766 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
2767 [Merlin Hughes <merlin@baltimore.ie>]
2769 *) Fix lots of warnings.
2770 [Richard Levitte <levitte@stacken.kth.se>]
2772 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
2773 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
2774 [Richard Levitte <levitte@stacken.kth.se>]
2776 *) Fix problems with sizeof(long) == 8.
2777 [Andy Polyakov <appro@fy.chalmers.se>]
2779 *) Change functions to ANSI C.
2782 *) Fix typos in error codes.
2783 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
2785 *) Remove defunct assembler files from Configure.
2788 *) SPARC v8 assembler BIGNUM implementation.
2789 [Andy Polyakov <appro@fy.chalmers.se>]
2791 *) Support for Certificate Policies extension: both print and set.
2792 Various additions to support the r2i method this uses.
2795 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
2796 return a const string when you are expecting an allocated buffer.
2799 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
2800 types DirectoryString and DisplayText.
2803 *) Add code to allow r2i extensions to access the configuration database,
2804 add an LHASH database driver and add several ctx helper functions.
2807 *) Fix an evil bug in bn_expand2() which caused various BN functions to
2808 fail when they extended the size of a BIGNUM.
2811 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
2812 support typesafe stack.
2815 *) Fix typo in SSL_[gs]et_options().
2816 [Nils Frostberg <nils@medcom.se>]
2818 *) Delete various functions and files that belonged to the (now obsolete)
2819 old X509V3 handling code.
2822 *) New Configure option "rsaref".
2825 *) Don't auto-generate pem.h.
2828 *) Introduce type-safe ASN.1 SETs.
2831 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
2832 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
2834 *) Introduce type-safe STACKs. This will almost certainly break lots of code
2835 that links with OpenSSL (well at least cause lots of warnings), but fear
2836 not: the conversion is trivial, and it eliminates loads of evil casts. A
2837 few STACKed things have been converted already. Feel free to convert more.
2838 In the fullness of time, I'll do away with the STACK type altogether.
2841 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
2842 specified in <certfile> by updating the entry in the index.txt file.
2843 This way one no longer has to edit the index.txt file manually for
2844 revoking a certificate. The -revoke option does the gory details now.
2845 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
2847 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
2848 `-text' option at all and this way the `-noout -text' combination was
2849 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
2850 [Ralf S. Engelschall]
2852 *) Make sure a corresponding plain text error message exists for the
2853 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
2854 verify callback function determined that a certificate was revoked.
2855 [Ralf S. Engelschall]
2857 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
2858 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
2859 all available cipers including rc5, which was forgotten until now.
2860 In order to let the testing shell script know which algorithms
2861 are available, a new (up to now undocumented) command
2862 "openssl list-cipher-commands" is used.
2865 *) Bugfix: s_client occasionally would sleep in select() when
2866 it should have checked SSL_pending() first.
2869 *) New functions DSA_do_sign and DSA_do_verify to provide access to
2870 the raw DSA values prior to ASN.1 encoding.
2873 *) Tweaks to Configure
2874 [Niels Poppe <niels@netbox.org>]
2876 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
2880 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
2883 *) New config option to avoid instructions that are illegal on the 80386.
2884 The default code is faster, but requires at least a 486.
2887 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
2888 SSL2_SERVER_VERSION (not used at all) macros, which are now the
2889 same as SSL2_VERSION anyway.
2892 *) New "-showcerts" option for s_client.
2895 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
2896 application. Various cleanups and fixes.
2899 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
2900 modify error routines to work internally. Add error codes and PBE init
2901 to library startup routines.
2904 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
2905 packing functions to asn1 and evp. Changed function names and error
2906 codes along the way.
2909 *) PKCS12 integration: and so it begins... First of several patches to
2910 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
2911 objects to objects.h
2914 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
2915 and display support for Thawte strong extranet extension.
2918 *) Add LinuxPPC support.
2919 [Jeff Dubrule <igor@pobox.org>]
2921 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
2922 bn_div_words in alpha.s.
2923 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
2925 *) Make sure the RSA OAEP test is skipped under -DRSAref because
2926 OAEP isn't supported when OpenSSL is built with RSAref.
2927 [Ulf Moeller <ulf@fitug.de>]
2929 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
2930 so they no longer are missing under -DNOPROTO.
2931 [Soren S. Jorvang <soren@t.dk>]
2934 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
2936 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
2937 doesn't work when the session is reused. Coming soon!
2940 *) Fix a security hole, that allows sessions to be reused in the wrong
2941 context thus bypassing client cert protection! All software that uses
2942 client certs and session caches in multiple contexts NEEDS PATCHING to
2943 allow session reuse! A fuller solution is in the works.
2944 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
2946 *) Some more source tree cleanups (removed obsolete files
2947 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
2948 permission on "config" script to be executable) and a fix for the INSTALL
2950 [Ulf Moeller <ulf@fitug.de>]
2952 *) Remove some legacy and erroneous uses of malloc, free instead of
2954 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
2956 *) Make rsa_oaep_test return non-zero on error.
2957 [Ulf Moeller <ulf@fitug.de>]
2959 *) Add support for native Solaris shared libraries. Configure
2960 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
2961 if someone would make that last step automatic.
2962 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
2964 *) ctx_size was not built with the right compiler during "make links". Fixed.
2967 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
2968 except NULL ciphers". This means the default cipher list will no longer
2969 enable NULL ciphers. They need to be specifically enabled e.g. with
2970 the string "DEFAULT:eNULL".
2973 *) Fix to RSA private encryption routines: if p < q then it would
2974 occasionally produce an invalid result. This will only happen with
2975 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
2978 *) Be less restrictive and allow also `perl util/perlpath.pl
2979 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
2980 because this way one can also use an interpreter named `perl5' (which is
2981 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
2982 installed as `perl').
2983 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
2985 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
2986 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
2988 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
2989 advapi32.lib to Win32 build and change the pem test comparision
2990 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
2991 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
2992 and crypto/des/ede_cbcm_enc.c.
2995 *) DES quad checksum was broken on big-endian architectures. Fixed.
2998 *) Comment out two functions in bio.h that aren't implemented. Fix up the
2999 Win32 test batch file so it (might) work again. The Win32 test batch file
3000 is horrible: I feel ill....
3003 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
3004 in e_os.h. Audit of header files to check ANSI and non ANSI
3005 sections: 10 functions were absent from non ANSI section and not exported
3006 from Windows DLLs. Fixed up libeay.num for new functions.
3009 *) Make `openssl version' output lines consistent.
3010 [Ralf S. Engelschall]
3012 *) Fix Win32 symbol export lists for BIO functions: Added
3013 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
3014 to ms/libeay{16,32}.def.
3015 [Ralf S. Engelschall]
3017 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
3018 fine under Unix and passes some trivial tests I've now added. But the
3019 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
3020 added to make sure no one expects that this stuff really works in the
3021 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
3022 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
3024 [Ralf S. Engelschall]
3026 *) Fix the generation of two part addresses in perl.
3027 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
3029 *) Add config entry for Linux on MIPS.
3030 [John Tobey <jtobey@channel1.com>]
3032 *) Make links whenever Configure is run, unless we are on Windoze.
3035 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
3036 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
3040 *) Add a useful kludge to allow package maintainers to specify compiler and
3041 other platforms details on the command line without having to patch the
3042 Configure script everytime: One now can use ``perl Configure
3043 <id>:<details>'', i.e. platform ids are allowed to have details appended
3044 to them (seperated by colons). This is treated as there would be a static
3045 pre-configured entry in Configure's %table under key <id> with value
3046 <details> and ``perl Configure <id>'' is called. So, when you want to
3047 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
3048 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
3049 now, which overrides the FreeBSD-elf entry on-the-fly.
3050 [Ralf S. Engelschall]
3052 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
3055 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
3056 on the `perl Configure ...' command line. This way one can compile
3057 OpenSSL libraries with Position Independent Code (PIC) which is needed
3058 for linking it into DSOs.
3059 [Ralf S. Engelschall]
3061 *) Remarkably, export ciphers were totally broken and no-one had noticed!
3065 *) Cleaned up the LICENSE document: The official contact for any license
3066 questions now is the OpenSSL core team under openssl-core@openssl.org.
3067 And add a paragraph about the dual-license situation to make sure people
3068 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
3069 to the OpenSSL toolkit.
3070 [Ralf S. Engelschall]
3072 *) General source tree makefile cleanups: Made `making xxx in yyy...'
3073 display consistent in the source tree and replaced `/bin/rm' by `rm'.
3074 Additonally cleaned up the `make links' target: Remove unnecessary
3075 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
3076 to speed processing and no longer clutter the display with confusing
3077 stuff. Instead only the actually done links are displayed.
3078 [Ralf S. Engelschall]
3080 *) Permit null encryption ciphersuites, used for authentication only. It used
3081 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
3082 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
3086 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
3087 signed attributes when verifying signatures (this would break them),
3088 the detached data encoding was wrong and public keys obtained using
3089 X509_get_pubkey() weren't freed.
3092 *) Add text documentation for the BUFFER functions. Also added a work around
3093 to a Win95 console bug. This was triggered by the password read stuff: the
3094 last character typed gets carried over to the next fread(). If you were
3095 generating a new cert request using 'req' for example then the last
3096 character of the passphrase would be CR which would then enter the first
3100 *) Added the new `Includes OpenSSL Cryptography Software' button as
3101 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
3102 button and can be used by applications based on OpenSSL to show the
3103 relationship to the OpenSSL project.
3104 [Ralf S. Engelschall]
3106 *) Remove confusing variables in function signatures in files
3107 ssl/ssl_lib.c and ssl/ssl.h.
3108 [Lennart Bong <lob@kulthea.stacken.kth.se>]
3110 *) Don't install bss_file.c under PREFIX/include/
3111 [Lennart Bong <lob@kulthea.stacken.kth.se>]
3113 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
3114 functions that return function pointers and has support for NT specific
3115 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
3116 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
3117 unsigned to signed types: this was killing the Win32 compile.
3120 *) Add new certificate file to stack functions,
3121 SSL_add_dir_cert_subjects_to_stack() and
3122 SSL_add_file_cert_subjects_to_stack(). These largely supplant
3123 SSL_load_client_CA_file(), and can be used to add multiple certs easily
3124 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
3125 This means that Apache-SSL and similar packages don't have to mess around
3126 to add as many CAs as they want to the preferred list.
3129 *) Experiment with doxygen documentation. Currently only partially applied to
3131 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
3132 openssl.doxy as the configuration file.
3135 *) Get rid of remaining C++-style comments which strict C compilers hate.
3136 [Ralf S. Engelschall, pointed out by Carlos Amengual]
3138 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
3139 compiled in by default: it has problems with large keys.
3142 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
3143 DH private keys and/or callback functions which directly correspond to
3144 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
3145 is needed for applications which have to configure certificates on a
3146 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
3148 For the RSA certificate situation is makes no difference, but
3149 for the DSA certificate situation this fixes the "no shared cipher"
3150 problem where the OpenSSL cipher selection procedure failed because the
3151 temporary keys were not overtaken from the context and the API provided
3152 no way to reconfigure them.
3153 The new functions now let applications reconfigure the stuff and they
3154 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
3155 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
3156 non-public-API function ssl_cert_instantiate() is used as a helper
3157 function and also to reduce code redundancy inside ssl_rsa.c.
3158 [Ralf S. Engelschall]
3160 *) Move s_server -dcert and -dkey options out of the undocumented feature
3161 area because they are useful for the DSA situation and should be
3162 recognized by the users.
3163 [Ralf S. Engelschall]
3165 *) Fix the cipher decision scheme for export ciphers: the export bits are
3166 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
3167 SSL_EXP_MASK. So, the original variable has to be used instead of the
3168 already masked variable.
3169 [Richard Levitte <levitte@stacken.kth.se>]
3171 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
3172 [Richard Levitte <levitte@stacken.kth.se>]
3174 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
3175 from `int' to `unsigned int' because it's a length and initialized by
3176 EVP_DigestFinal() which expects an `unsigned int *'.
3177 [Richard Levitte <levitte@stacken.kth.se>]
3179 *) Don't hard-code path to Perl interpreter on shebang line of Configure
3180 script. Instead use the usual Shell->Perl transition trick.
3181 [Ralf S. Engelschall]
3183 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
3184 (in addition to RSA certificates) to match the behaviour of `openssl dsa
3185 -noout -modulus' as it's already the case for `openssl rsa -noout
3186 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
3187 currently the public key is printed (a decision which was already done by
3188 `openssl dsa -modulus' in the past) which serves a similar purpose.
3189 Additionally the NO_RSA no longer completely removes the whole -modulus
3190 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
3192 [Ralf S. Engelschall]
3194 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
3195 BIO. See the source (crypto/evp/bio_ok.c) for more info.
3196 [Arne Ansper <arne@ats.cyber.ee>]
3198 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
3199 to be added. Now both 'req' and 'ca' can use new objects defined in the
3203 *) Add cool BIO that does syslog (or event log on NT).
3204 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
3206 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
3207 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
3208 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
3209 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
3212 *) Add preliminary config info for new extension code.
3215 *) Make RSA_NO_PADDING really use no padding.
3216 [Ulf Moeller <ulf@fitug.de>]
3218 *) Generate errors when private/public key check is done.
3221 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
3222 for some CRL extensions and new objects added.
3225 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
3226 key usage extension and fuller support for authority key id.
3229 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
3230 padding method for RSA, which is recommended for new applications in PKCS
3231 #1 v2.0 (RFC 2437, October 1998).
3232 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
3233 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
3234 against Bleichbacher's attack on RSA.
3235 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
3238 *) Updates to the new SSL compression code
3239 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
3241 *) Fix so that the version number in the master secret, when passed
3242 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
3243 (because the server will not accept higher), that the version number
3244 is 0x03,0x01, not 0x03,0x00
3245 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
3247 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
3248 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
3249 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
3252 *) Support for RAW extensions where an arbitrary extension can be
3253 created by including its DER encoding. See apps/openssl.cnf for
3257 *) Make sure latest Perl versions don't interpret some generated C array
3258 code as Perl array code in the crypto/err/err_genc.pl script.
3259 [Lars Weber <3weber@informatik.uni-hamburg.de>]
3261 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
3262 not many people have the assembler. Various Win32 compilation fixes and
3263 update to the INSTALL.W32 file with (hopefully) more accurate Win32
3267 *) Modify configure script 'Configure' to automatically create crypto/date.h
3268 file under Win32 and also build pem.h from pem.org. New script
3269 util/mkfiles.pl to create the MINFO file on environments that can't do a
3270 'make files': perl util/mkfiles.pl >MINFO should work.
3273 *) Major rework of DES function declarations, in the pursuit of correctness
3274 and purity. As a result, many evil casts evaporated, and some weirdness,
3275 too. You may find this causes warnings in your code. Zapping your evil
3276 casts will probably fix them. Mostly.
3279 *) Fix for a typo in asn1.h. Bug fix to object creation script
3280 obj_dat.pl. It considered a zero in an object definition to mean
3281 "end of object": none of the objects in objects.h have any zeros
3282 so it wasn't spotted.
3283 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
3285 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
3286 Masking (CBCM). In the absence of test vectors, the best I have been able
3287 to do is check that the decrypt undoes the encrypt, so far. Send me test
3288 vectors if you have them.
3291 *) Correct calculation of key length for export ciphers (too much space was
3292 allocated for null ciphers). This has not been tested!
3295 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
3296 message is now correct (it understands "crypto" and "ssl" on its
3297 command line). There is also now an "update" option. This will update
3298 the util/ssleay.num and util/libeay.num files with any new functions.
3300 perl util/mkdef.pl crypto ssl update
3301 it will update them.
3304 *) Overhauled the Perl interface (perl/*):
3305 - ported BN stuff to OpenSSL's different BN library
3306 - made the perl/ source tree CVS-aware
3307 - renamed the package from SSLeay to OpenSSL (the files still contain
3308 their history because I've copied them in the repository)
3309 - removed obsolete files (the test scripts will be replaced
3310 by better Test::Harness variants in the future)
3311 [Ralf S. Engelschall]
3313 *) First cut for a very conservative source tree cleanup:
3314 1. merge various obsolete readme texts into doc/ssleay.txt
3315 where we collect the old documents and readme texts.
3316 2. remove the first part of files where I'm already sure that we no
3317 longer need them because of three reasons: either they are just temporary
3318 files which were left by Eric or they are preserved original files where
3319 I've verified that the diff is also available in the CVS via "cvs diff
3320 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
3321 the crypto/md/ stuff).
3322 [Ralf S. Engelschall]
3324 *) More extension code. Incomplete support for subject and issuer alt
3325 name, issuer and authority key id. Change the i2v function parameters
3326 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
3327 what that's for :-) Fix to ASN1 macro which messed up
3328 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
3331 *) Preliminary support for ENUMERATED type. This is largely copied from the
3335 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
3336 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
3338 *) Make sure `make rehash' target really finds the `openssl' program.
3339 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
3341 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
3342 like to hear about it if this slows down other processors.
3345 *) Add CygWin32 platform information to Configure script.
3346 [Alan Batie <batie@aahz.jf.intel.com>]
3348 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
3349 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
3351 *) New program nseq to manipulate netscape certificate sequences
3354 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
3358 *) Fixes to BN code. Previously the default was to define BN_RECURSION
3359 but the BN code had some problems that would cause failures when
3360 doing certificate verification and some other functions.
3361 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
3363 *) Add ASN1 and PEM code to support netscape certificate sequences.
3366 *) Add ASN1 and PEM code to support netscape certificate sequences.
3369 *) Add several PKIX and private extended key usage OIDs.
3372 *) Modify the 'ca' program to handle the new extension code. Modify
3373 openssl.cnf for new extension format, add comments.
3376 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
3377 and add a sample to openssl.cnf so req -x509 now adds appropriate
3381 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
3382 error code, add initial support to X509_print() and x509 application.
3385 *) Takes a deep breath and start addding X509 V3 extension support code. Add
3386 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
3387 stuff is currently isolated and isn't even compiled yet.
3390 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
3391 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
3392 Removed the versions check from X509 routines when loading extensions:
3393 this allows certain broken certificates that don't set the version
3394 properly to be processed.
3397 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
3398 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
3399 can still be regenerated with "make depend".
3402 *) Spelling mistake in C version of CAST-128.
3403 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
3405 *) Changes to the error generation code. The perl script err-code.pl
3406 now reads in the old error codes and retains the old numbers, only
3407 adding new ones if necessary. It also only changes the .err files if new
3408 codes are added. The makefiles have been modified to only insert errors
3409 when needed (to avoid needlessly modifying header files). This is done
3410 by only inserting errors if the .err file is newer than the auto generated
3411 C file. To rebuild all the error codes from scratch (the old behaviour)
3412 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
3413 or delete all the .err files.
3416 *) CAST-128 was incorrectly implemented for short keys. The C version has
3417 been fixed, but is untested. The assembler versions are also fixed, but
3418 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
3419 to regenerate it if needed.
3420 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
3421 Hagino <itojun@kame.net>]
3423 *) File was opened incorrectly in randfile.c.
3424 [Ulf Möller <ulf@fitug.de>]
3426 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
3427 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
3428 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
3429 al: it's just almost always a UTCTime. Note this patch adds new error
3430 codes so do a "make errors" if there are problems.
3433 *) Correct Linux 1 recognition in config.
3434 [Ulf Möller <ulf@fitug.de>]
3436 *) Remove pointless MD5 hash when using DSA keys in ca.
3437 [Anonymous <nobody@replay.com>]
3439 *) Generate an error if given an empty string as a cert directory. Also
3440 generate an error if handed NULL (previously returned 0 to indicate an
3441 error, but didn't set one).
3442 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
3444 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
3447 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
3448 parameters. This was causing a warning which killed off the Win32 compile.
3451 *) Remove C++ style comments from crypto/bn/bn_local.h.
3452 [Neil Costigan <neil.costigan@celocom.com>]
3454 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
3455 based on a text string, looking up short and long names and finally
3456 "dot" format. The "dot" format stuff didn't work. Added new function
3457 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
3458 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
3459 OID is not part of the table.
3462 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
3463 X509_LOOKUP_by_alias().
3466 *) Sort openssl functions by name.
3469 *) Get the gendsa program working (hopefully) and add it to app list. Remove
3470 encryption from sample DSA keys (in case anyone is interested the password
3474 *) Make _all_ *_free functions accept a NULL pointer.
3475 [Frans Heymans <fheymans@isaserver.be>]
3477 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
3479 [Anonymous <nobody@replay.com>]
3481 *) s_server should send the CAfile as acceptable CAs, not its own cert.
3482 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
3484 *) Don't blow it for numeric -newkey arguments to apps/req.
3485 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
3487 *) Temp key "for export" tests were wrong in s3_srvr.c.
3488 [Anonymous <nobody@replay.com>]
3490 *) Add prototype for temp key callback functions
3491 SSL_CTX_set_tmp_{rsa,dh}_callback().
3494 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
3495 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
3498 *) X509_name_add_entry() freed the wrong thing after an error.
3499 [Arne Ansper <arne@ats.cyber.ee>]
3501 *) rsa_eay.c would attempt to free a NULL context.
3502 [Arne Ansper <arne@ats.cyber.ee>]
3504 *) BIO_s_socket() had a broken should_retry() on Windoze.
3505 [Arne Ansper <arne@ats.cyber.ee>]
3507 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
3508 [Arne Ansper <arne@ats.cyber.ee>]
3510 *) Make sure the already existing X509_STORE->depth variable is initialized
3511 in X509_STORE_new(), but document the fact that this variable is still
3512 unused in the certificate verification process.
3513 [Ralf S. Engelschall]
3515 *) Fix the various library and apps files to free up pkeys obtained from
3516 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
3519 *) Fix reference counting in X509_PUBKEY_get(). This makes
3520 demos/maurice/example2.c work, amongst others, probably.
3521 [Steve Henson and Ben Laurie]
3523 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
3524 `openssl' and second, the shortcut symlinks for the `openssl <command>'
3525 are no longer created. This way we have a single and consistent command
3526 line interface `openssl <command>', similar to `cvs <command>'.
3527 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
3529 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
3530 BIT STRING wrapper always have zero unused bits.
3533 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
3536 *) Make the top-level INSTALL documentation easier to understand.
3539 *) Makefiles updated to exit if an error occurs in a sub-directory
3540 make (including if user presses ^C) [Paul Sutton]
3542 *) Make Montgomery context stuff explicit in RSA data structure.
3545 *) Fix build order of pem and err to allow for generated pem.h.
3548 *) Fix renumbering bug in X509_NAME_delete_entry().
3551 *) Enhanced the err-ins.pl script so it makes the error library number
3552 global and can add a library name. This is needed for external ASN1 and
3553 other error libraries.
3556 *) Fixed sk_insert which never worked properly.
3559 *) Fix ASN1 macros so they can handle indefinite length construted
3560 EXPLICIT tags. Some non standard certificates use these: they can now
3564 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
3565 into a single doc/ssleay.txt bundle. This way the information is still
3566 preserved but no longer messes up this directory. Now it's new room for
3567 the new set of documenation files.
3568 [Ralf S. Engelschall]
3570 *) SETs were incorrectly DER encoded. This was a major pain, because they
3571 shared code with SEQUENCEs, which aren't coded the same. This means that
3572 almost everything to do with SETs or SEQUENCEs has either changed name or
3573 number of arguments.
3574 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
3576 *) Fix test data to work with the above.
3579 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
3580 was already fixed by Eric for 0.9.1 it seems.
3581 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
3583 *) Autodetect FreeBSD3.
3586 *) Fix various bugs in Configure. This affects the following platforms:
3590 unixware-2.0-pentium
3594 *) Eliminate generated files from CVS. Reorder tests to regenerate files
3595 before they are needed.
3598 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
3602 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
3604 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
3605 changed SSLeay to OpenSSL in version strings.
3606 [Ralf S. Engelschall]
3608 *) Some fixups to the top-level documents.
3611 *) Fixed the nasty bug where rsaref.h was not found under compile-time
3612 because the symlink to include/ was missing.
3613 [Ralf S. Engelschall]
3615 *) Incorporated the popular no-RSA/DSA-only patches
3616 which allow to compile a RSA-free SSLeay.
3617 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
3619 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
3620 when "ssleay" is still not found.
3621 [Ralf S. Engelschall]
3623 *) Added more platforms to Configure: Cray T3E, HPUX 11,
3624 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
3626 *) Updated the README file.
3627 [Ralf S. Engelschall]
3629 *) Added various .cvsignore files in the CVS repository subdirs
3630 to make a "cvs update" really silent.
3631 [Ralf S. Engelschall]
3633 *) Recompiled the error-definition header files and added
3634 missing symbols to the Win32 linker tables.
3635 [Ralf S. Engelschall]
3637 *) Cleaned up the top-level documents;
3638 o new files: CHANGES and LICENSE
3639 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
3640 o merged COPYRIGHT into LICENSE
3641 o removed obsolete TODO file
3642 o renamed MICROSOFT to INSTALL.W32
3643 [Ralf S. Engelschall]
3645 *) Removed dummy files from the 0.9.1b source tree:
3646 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
3647 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
3648 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
3649 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
3650 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
3651 [Ralf S. Engelschall]
3653 *) Added various platform portability fixes.
3656 *) The Genesis of the OpenSSL rpject:
3657 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
3658 Young and Tim J. Hudson created while they were working for C2Net until
3660 [The OpenSSL Project]
3663 Changes between 0.9.0b and 0.9.1b [not released]
3665 *) Updated a few CA certificates under certs/
3668 *) Changed some BIGNUM api stuff.
3671 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
3672 DGUX x86, Linux Alpha, etc.
3675 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
3676 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
3680 *) Add -strparse option to asn1pars program which parses nested
3682 [Dr Stephen Henson <shenson@bigfoot.com>]
3684 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
3687 *) DSA fix for "ca" program.
3690 *) Added "-genkey" option to "dsaparam" program.
3693 *) Added RIPE MD160 (rmd160) message digest.
3696 *) Added -a (all) option to "ssleay version" command.
3699 *) Added PLATFORM define which is the id given to Configure.
3702 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
3705 *) Extended the ASN.1 parser routines.
3708 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
3711 *) Added a BN_CTX to the BN library.
3714 *) Fixed the weak key values in DES library
3717 *) Changed API in EVP library for cipher aliases.
3720 *) Added support for RC2/64bit cipher.
3723 *) Converted the lhash library to the crypto/mem.c functions.
3726 *) Added more recognized ASN.1 object ids.
3729 *) Added more RSA padding checks for SSL/TLS.
3732 *) Added BIO proxy/filter functionality.
3735 *) Added extra_certs to SSL_CTX which can be used
3736 send extra CA certificates to the client in the CA cert chain sending
3737 process. It can be configured with SSL_CTX_add_extra_chain_cert().
3740 *) Now Fortezza is denied in the authentication phase because
3741 this is key exchange mechanism is not supported by SSLeay at all.
3744 *) Additional PKCS1 checks.
3747 *) Support the string "TLSv1" for all TLS v1 ciphers.
3750 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
3751 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
3754 *) Fixed a few memory leaks.
3757 *) Fixed various code and comment typos.
3760 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
3761 bytes sent in the client random.
3762 [Edward Bishop <ebishop@spyglass.com>]