4 Changes between 0.9.6 and 0.9.7 [xx XXX 2000]
6 *) Make BN_mod_inverse faster by explicitly handling small quotients
7 in the Euclid loop instead of always using BN_div.
8 (Speed gain about 20% for small moduli [256 or 512 bits], about
9 30% for larger ones [1024 or 2048 bits].)
12 *) Disable ssl2_peek and ssl3_peek (i.e., both implementations
13 of SSL_peek) because they both are completely broken.
14 They will be fixed RSN by adding an additional 'peek' parameter
15 to the internal read functions.
18 *) New function BN_kronecker.
21 *) Fix BN_gcd so that it works on negative inputs; the result is
22 positive unless both parameters are zero.
23 Previously something reasonably close to an infinite loop was
24 possible because numbers could be growing instead of shrinking
25 in the implementation of Euclid's algorithm.
28 *) Fix BN_is_word() and BN_is_one() macros to take into account the
29 sign of the number in question.
31 Fix BN_is_word(a,w) to work correctly for w == 0.
33 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
34 because its test if the absolute value of 'a' equals 'w'.
35 Note that BN_abs_is_word does *not* handle w == 0 reliably;
36 it exists mostly for use in the implementations of BN_is_zero(),
37 BN_is_one(), and BN_is_word().
40 *) Initialise "ex_data" member of an RSA structure prior to calling the
41 method-specific "init()" handler, and clean up ex_data after calling
42 the method-specific "finish()" handler. Previously, this was happening
46 *) New function BN_swap.
49 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
50 the exponentiation functions are more likely to produce reasonable
51 results on negative inputs.
54 *) Change BN_mod_mul so that the result is always non-negative.
55 Previously, it could be negative if one of the factors was negative;
56 I don't think anyone really wanted that behaviour.
59 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
60 (except for exponentation, which stays in crypto/bn/bn_exp.c,
61 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
62 and add new functions:
75 These functions always generate non-negative results.
77 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
78 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
80 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
81 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
83 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
85 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
86 was actually never needed) and in BN_mul(). The removal in BN_mul()
87 required a small change in bn_mul_part_recursive() and the addition
88 of the static functions bn_cmp_part_words(), bn_sub_part_words()
89 and bn_add_part_words() which do the same thing as bn_cmp_words(),
90 bn_sub_words() and bn_add_words() except they take arrays with
94 *) In 'openssl passwd', verify passwords read from the terminal
95 unless the '-salt' option is used (which usually means that
96 verification would just waste user's time since the resulting
97 hash is going to be compared with some given password hash)
98 or the new '-noverify' option is used.
100 This is an incompatible change, but it does not affect
101 non-interactive use of 'openssl passwd' (passwords on the command
102 line, '-stdin' option, '-in ...' option) and thus should not
106 *) Remove all references to RSAref, since there's no more need for it.
109 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
110 The previous value, 12, was not always sufficient for BN_mod_exp().
113 *) Make DSO load along a path given through an environment variable
114 (SHLIB_PATH) with shl_load().
117 *) Constify the ENGINE code as a result of BIGNUM constification.
118 Also constify the RSA code and most things related to it. In a
119 few places, most notable in the depth of the ASN.1 code, ugly
120 casts back to non-const were required (to be solved at a later
124 *) Make it so the openssl application has all engines loaded by default.
127 *) Constify the BIGNUM routines a little more.
130 *) Make sure that shared libraries get the internal name engine with
131 the full version number and not just 0. This should mark the
132 shared libraries as not backward compatible. Of course, this should
133 be changed again when we can guarantee backward binary compatibility.
136 *) Add the following functions:
142 ENGINE_load_builtin_engines()
144 That way, an application can itself choose if external engines that
145 are built-in in OpenSSL shall ever be used or not. The benefit is
146 that applications won't have to be linked with libdl or other dso
147 libraries unless it's really needed.
149 Changed 'openssl engine' to load all engines on demand.
150 Changed the engine header files to avoid the duplication of some
151 declarations (they differed!).
154 *) 'openssl engine' can now list capabilities.
157 *) Better error reporting in 'openssl engine'.
160 *) Never call load_dh_param(NULL) in s_server.
163 *) Add engine application. It can currently list engines by name and
164 identity, and test if they are actually available.
167 *) Add support for shared libraries under Irix.
168 [Albert Chin-A-Young <china@thewrittenword.com>]
170 *) Improve RPM specification file by forcing symbolic linking and making
171 sure the installed documentation is also owned by root.root.
172 [Damien Miller <djm@mindrot.org>]
174 *) Add configuration option to build on Linux on both big-endian and
176 [Ralf Baechle <ralf@uni-koblenz.de>]
178 *) Give the OpenSSL applications more possibilities to make use of
179 keys (public as well as private) handled by engines.
182 *) Add OCSP code that comes from CertCo.
185 *) Add VMS support for the Rijndael code.
188 *) Added untested support for Nuron crypto accelerator.
191 *) Add support for external cryptographic devices. This code was
192 previously distributed separately as the "engine" branch.
193 [Geoff Thorpe, Richard Levitte]
195 *) Rework the filename-translation in the DSO code. It is now possible to
196 have far greater control over how a "name" is turned into a filename
197 depending on the operating environment and any oddities about the
198 different shared library filenames on each system.
201 *) Support threads on FreeBSD-elf in Configure.
204 *) Add the possibility to create shared libraries on HP-UX
207 *) Fix for SHA1 assembly problem with MASM: it produces
208 warnings about corrupt line number information when assembling
209 with debugging information. This is caused by the overlapping
211 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
214 NCONF_get_number() has no error checking at all. As a replacement,
215 NCONF_get_number_e() is defined (_e for "error checking") and is
216 promoted strongly. The old NCONF_get_number is kept around for
217 binary backward compatibility.
218 Make it possible for methods to load from something other than a BIO,
219 by providing a function pointer that is given a name instead of a BIO.
220 For example, this could be used to load configuration data from an
224 *) Fix typo in get_cert_by_subject() in by_dir.c
225 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
227 *) Rework the system to generate shared libraries:
229 - Make note of the expected extension for the shared libraries and
230 if there is a need for symbolic links from for example libcrypto.so.0
231 to libcrypto.so.0.9.7. There is extended info in Configure for
234 - Make as few rebuilds of the shared libraries as possible.
236 - Still avoid linking the OpenSSL programs with the shared libraries.
238 - When installing, install the shared libraries separately from the
242 *) Fix for non blocking accept BIOs. Added new I/O special reason
243 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
244 with non blocking I/O was not possible because no retry code was
245 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
249 *) Added the beginnings of Rijndael support.
252 *) Fix for bug in DirectoryString mask setting. Add support for
253 X509_NAME_print_ex() in 'req' and X509_print_ex() function
254 to allow certificate printing to more controllable, additional
255 'certopt' option to 'x509' to allow new printing options to be
259 *) Clean old EAY MD5 hack from e_os.h.
262 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
264 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
265 and not in SSL_clear because the latter is also used by the
266 accept/connect functions; previously, the settings made by
267 SSL_set_read_ahead would be lost during the handshake.
268 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
270 *) Correct util/mkdef.pl to be selective about disabled algorithms.
271 Previously, it would create entries for disableed algorithms no
275 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
277 *) In ssl23_get_client_hello, generate an error message when faced
278 with an initial SSL 3.0/TLS record that is too small to contain the
279 first two bytes of the ClientHello message, i.e. client_version.
280 (Note that this is a pathologic case that probably has never happened
281 in real life.) The previous approach was to use the version number
282 from the record header as a substitute; but our protocol choice
283 should not depend on that one because it is not authenticated
284 by the Finished messages.
287 *) More robust randomness gathering functions for Windows.
288 [Jeffrey Altman <jaltman@columbia.edu>]
290 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
291 not set then we don't setup the error code for issuer check errors
292 to avoid possibly overwriting other errors which the callback does
293 handle. If an application does set the flag then we assume it knows
294 what it is doing and can handle the new informational codes
298 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
299 a general "ANY" type, as such it should be able to decode anything
300 including tagged types. However it didn't check the class so it would
301 wrongly interpret tagged types in the same way as their universal
302 counterpart and unknown types were just rejected. Changed so that the
303 tagged and unknown types are handled in the same way as a SEQUENCE:
304 that is the encoding is stored intact. There is also a new type
305 "V_ASN1_OTHER" which is used when the class is not universal, in this
306 case we have no idea what the actual type is so we just lump them all
310 *) On VMS, stdout may very well lead to a file that is written to
311 in a record-oriented fashion. That means that every write() will
312 write a separate record, which will be read separately by the
313 programs trying to read from it. This can be very confusing.
315 The solution is to put a BIO filter in the way that will buffer
316 text until a linefeed is reached, and then write everything a
317 line at a time, so every record written will be an actual line,
318 not chunks of lines and not (usually doesn't happen, but I've
319 seen it once) several lines in one record. BIO_f_linebuffer() is
322 Currently, it's a VMS-only method, because that's where it has
323 been tested well enough.
326 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
327 it can return incorrect results.
328 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
329 but it was in 0.9.6-beta[12].)
332 *) Disable the check for content being present when verifying detached
333 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
334 include zero length content when signing messages.
337 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
338 BIO_ctrl (for BIO pairs).
341 *) Add DSO method for VMS.
344 *) Bug fix: Montgomery multiplication could produce results with the
348 *) Add RPM specification openssl.spec and modify it to build three
349 packages. The default package contains applications, application
350 documentation and run-time libraries. The devel package contains
351 include files, static libraries and function documentation. The
352 doc package contains the contents of the doc directory. The original
353 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
356 *) Add a large number of documentation files for many SSL routines.
357 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
359 *) Add a configuration entry for Sony News 4.
360 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
362 *) Don't set the two most significant bits to one when generating a
363 random number < q in the DSA library.
366 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
367 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
368 the underlying transport is blocking) if a handshake took place.
369 (The default behaviour is needed by applications such as s_client
370 and s_server that use select() to determine when to use SSL_read;
371 but for applications that know in advance when to expect data, it
372 just makes things more complicated.)
375 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
379 *) Add a few more EBCDIC conditionals that make `req' and `x509'
380 work better on such systems.
381 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
383 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
384 Update PKCS12_parse() so it copies the friendlyName and the
385 keyid to the certificates aux info.
388 *) Fix bug in PKCS7_verify() which caused an infinite loop
389 if there was more than one signature.
390 [Sven Uszpelkat <su@celocom.de>]
392 *) Major change in util/mkdef.pl to include extra information
393 about each symbol, as well as presentig variables as well
394 as functions. This change means that there's n more need
395 to rebuild the .num files when some algorithms are excluded.
398 *) Allow the verify time to be set by an application,
399 rather than always using the current time.
402 *) Phase 2 verify code reorganisation. The certificate
403 verify code now looks up an issuer certificate by a
404 number of criteria: subject name, authority key id
405 and key usage. It also verifies self signed certificates
406 by the same criteria. The main comparison function is
407 X509_check_issued() which performs these checks.
409 Lot of changes were necessary in order to support this
410 without completely rewriting the lookup code.
412 Authority and subject key identifier are now cached.
414 The LHASH 'certs' is X509_STORE has now been replaced
415 by a STACK_OF(X509_OBJECT). This is mainly because an
416 LHASH can't store or retrieve multiple objects with
419 As a result various functions (which were all internal
420 use only) have changed to handle the new X509_STORE
421 structure. This will break anything that messed round
422 with X509_STORE internally.
424 The functions X509_STORE_add_cert() now checks for an
425 exact match, rather than just subject name.
427 The X509_STORE API doesn't directly support the retrieval
428 of multiple certificates matching a given criteria, however
429 this can be worked round by performing a lookup first
430 (which will fill the cache with candidate certificates)
431 and then examining the cache for matches. This is probably
432 the best we can do without throwing out X509_LOOKUP
433 entirely (maybe later...).
435 The X509_VERIFY_CTX structure has been enhanced considerably.
437 All certificate lookup operations now go via a get_issuer()
438 callback. Although this currently uses an X509_STORE it
439 can be replaced by custom lookups. This is a simple way
440 to bypass the X509_STORE hackery necessary to make this
441 work and makes it possible to use more efficient techniques
442 in future. A very simple version which uses a simple
443 STACK for its trusted certificate store is also provided
444 using X509_STORE_CTX_trusted_stack().
446 The verify_cb() and verify() callbacks now have equivalents
447 in the X509_STORE_CTX structure.
449 X509_STORE_CTX also has a 'flags' field which can be used
450 to customise the verify behaviour.
453 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
454 excludes S/MIME capabilities.
457 *) When a certificate request is read in keep a copy of the
458 original encoding of the signed data and use it when outputing
459 again. Signatures then use the original encoding rather than
460 a decoded, encoded version which may cause problems if the
461 request is improperly encoded.
464 *) For consistency with other BIO_puts implementations, call
465 buffer_write(b, ...) directly in buffer_puts instead of calling
468 In BIO_puts, increment b->num_write as in BIO_write.
469 [Peter.Sylvester@EdelWeb.fr]
471 *) Fix BN_mul_word for the case where the word is 0. (We have to use
472 BN_zero, we may not return a BIGNUM with an array consisting of
476 *) Avoid calling abort() from within the library when problems are
477 detected, except if preprocessor symbols have been defined
478 (such as REF_CHECK, BN_DEBUG etc.).
481 *) New openssl application 'rsautl'. This utility can be
482 used for low level RSA operations. DER public key
483 BIO/fp routines also added.
486 *) New Configure entry and patches for compiling on QNX 4.
487 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
489 *) A demo state-machine implementation was sponsored by
490 Nuron (http://www.nuron.com/) and is now available in
494 *) New options added to the 'dgst' utility for signature
495 generation and verification.
498 *) Unrecognized PKCS#7 content types are now handled via a
499 catch all ASN1_TYPE structure. This allows unsupported
500 types to be stored as a "blob" and an application can
501 encode and decode it manually.
504 *) Fix various signed/unsigned issues to make a_strex.c
506 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
508 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
509 length if passed a buffer. ASN1_INTEGER_to_BN failed
510 if passed a NULL BN and its argument was negative.
511 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
513 *) Modification to PKCS#7 encoding routines to output definite
514 length encoding. Since currently the whole structures are in
515 memory there's not real point in using indefinite length
516 constructed encoding. However if OpenSSL is compiled with
517 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
520 *) Added BIO_vprintf() and BIO_vsnprintf().
523 *) Added more prefixes to parse for in the the strings written
524 through a logging bio, to cover all the levels that are available
525 through syslog. The prefixes are now:
527 PANIC, EMERG, EMR => LOG_EMERG
528 ALERT, ALR => LOG_ALERT
529 CRIT, CRI => LOG_CRIT
530 ERROR, ERR => LOG_ERR
531 WARNING, WARN, WAR => LOG_WARNING
532 NOTICE, NOTE, NOT => LOG_NOTICE
533 INFO, INF => LOG_INFO
534 DEBUG, DBG => LOG_DEBUG
536 and as before, if none of those prefixes are present at the
537 beginning of the string, LOG_ERR is chosen.
539 On Win32, the LOG_* levels are mapped according to this:
541 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
542 LOG_WARNING => EVENTLOG_WARNING_TYPE
543 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
547 *) Made it possible to reconfigure with just the configuration
548 argument "reconf" or "reconfigure". The command line arguments
549 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
550 and are retrieved from there when reconfiguring.
554 [Assar Westerlund <assar@sics.se>, Richard Levitte]
556 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
559 *) The obj_dat.pl script was messing up the sorting of object
560 names. The reason was that it compared the quoted version
561 of strings as a result "OCSP" > "OCSP Signing" because
562 " > SPACE. Changed script to store unquoted versions of
563 names and add quotes on output. It was also omitting some
564 names from the lookup table if they were given a default
565 value (that is if SN is missing it is given the same
566 value as LN and vice versa), these are now added on the
567 grounds that if an object has a name we should be able to
568 look it up. Finally added warning output when duplicate
569 short or long names are found.
572 *) Changes needed for Tandem NSK.
573 [Scott Uroff <scott@xypro.com>]
575 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
576 RSA_padding_check_SSLv23(), special padding was never detected
577 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
578 version rollback attacks was not effective.
580 In s23_clnt.c, don't use special rollback-attack detection padding
581 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
582 client; similarly, in s23_srvr.c, don't do the rollback check if
583 SSL 2.0 is the only protocol enabled in the server.
586 *) Make it possible to get hexdumps of unprintable data with 'openssl
587 asn1parse'. By implication, the functions ASN1_parse_dump() and
588 BIO_dump_indent() are added.
591 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
592 these print out strings and name structures based on various
593 flags including RFC2253 support and proper handling of
594 multibyte characters. Added options to the 'x509' utility
595 to allow the various flags to be set.
598 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
599 Also change the functions X509_cmp_current_time() and
600 X509_gmtime_adj() work with an ASN1_TIME structure,
601 this will enable certificates using GeneralizedTime in validity
605 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
606 negative public key encodings) on by default,
607 NO_NEG_PUBKEY_BUG can be set to disable it.
610 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
611 content octets. An i2c_ASN1_OBJECT is unnecessary because
612 the encoding can be trivially obtained from the structure.
615 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
616 not read locks (CRYPTO_r_[un]lock).
619 *) A first attempt at creating official support for shared
620 libraries through configuration. I've kept it so the
621 default is static libraries only, and the OpenSSL programs
622 are always statically linked for now, but there are
623 preparations for dynamic linking in place.
624 This has been tested on Linux and True64.
627 *) Randomness polling function for Win9x, as described in:
628 Peter Gutmann, Software Generation of Practically Strong
632 *) Fix so PRNG is seeded in req if using an already existing
636 *) New options to smime application. -inform and -outform
637 allow alternative formats for the S/MIME message including
638 PEM and DER. The -content option allows the content to be
639 specified separately. This should allow things like Netscape
640 form signing output easier to verify.
643 *) Fix the ASN1 encoding of tags using the 'long form'.
646 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
647 STRING types. These convert content octets to and from the
648 underlying type. The actual tag and length octets are
649 already assumed to have been read in and checked. These
650 are needed because all other string types have virtually
651 identical handling apart from the tag. By having versions
652 of the ASN1 functions that just operate on content octets
653 IMPLICIT tagging can be handled properly. It also allows
654 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
655 and ASN1_INTEGER are identical apart from the tag.
658 *) Change the handling of OID objects as follows:
660 - New object identifiers are inserted in objects.txt, following
661 the syntax given in objects.README.
662 - objects.pl is used to process obj_mac.num and create a new
664 - obj_dat.pl is used to create a new obj_dat.h, using the data in
667 This is currently kind of a hack, and the perl code in objects.pl
668 isn't very elegant, but it works as I intended. The simplest way
669 to check that it worked correctly is to look in obj_dat.h and
670 check the array nid_objs and make sure the objects haven't moved
671 around (this is important!). Additions are OK, as well as
672 consistent name changes.
675 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
678 *) Addition of the command line parameter '-rand file' to 'openssl req'.
679 The given file adds to whatever has already been seeded into the
680 random pool through the RANDFILE configuration file option or
681 environment variable, or the default random state file.
684 *) mkstack.pl now sorts each macro group into lexical order.
685 Previously the output order depended on the order the files
686 appeared in the directory, resulting in needless rewriting
690 *) Patches to make OpenSSL compile under Win32 again. Mostly
691 work arounds for the VC++ problem that it treats func() as
692 func(void). Also stripped out the parts of mkdef.pl that
693 added extra typesafe functions: these no longer exist.
696 *) Reorganisation of the stack code. The macros are now all
697 collected in safestack.h . Each macro is defined in terms of
698 a "stack macro" of the form SKM_<name>(type, a, b). The
699 DEBUG_SAFESTACK is now handled in terms of function casts,
700 this has the advantage of retaining type safety without the
701 use of additional functions. If DEBUG_SAFESTACK is not defined
702 then the non typesafe macros are used instead. Also modified the
703 mkstack.pl script to handle the new form. Needs testing to see
704 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
705 the default if no major problems. Similar behaviour for ASN1_SET_OF
709 *) When some versions of IIS use the 'NET' form of private key the
710 key derivation algorithm is different. Normally MD5(password) is
711 used as a 128 bit RC4 key. In the modified case
712 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
713 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
714 as the old Netscape_RSA functions except they have an additional
715 'sgckey' parameter which uses the modified algorithm. Also added
716 an -sgckey command line option to the rsa utility. Thanks to
717 Adrian Peck <bertie@ncipher.com> for posting details of the modified
718 algorithm to openssl-dev.
721 *) The evp_local.h macros were using 'c.##kname' which resulted in
722 invalid expansion on some systems (SCO 5.0.5 for example).
723 Corrected to 'c.kname'.
724 [Phillip Porch <root@theporch.com>]
726 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
727 a STACK of email addresses from a certificate or request, these look
728 in the subject name and the subject alternative name extensions and
729 omit any duplicate addresses.
732 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
733 This makes DSA verification about 2 % faster.
736 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
737 (meaning that now 2^5 values will be precomputed, which is only 4 KB
738 plus overhead for 1024 bit moduli).
739 This makes exponentiations about 0.5 % faster for 1024 bit
740 exponents (as measured by "openssl speed rsa2048").
743 *) Rename memory handling macros to avoid conflicts with other
745 Malloc => OPENSSL_malloc
746 Malloc_locked => OPENSSL_malloc_locked
747 Realloc => OPENSSL_realloc
751 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
752 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
756 [John Jarvie <jjarvie@newsguy.com>]
758 *) The type-safe stack code has been rejigged. It is now only compiled
759 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
760 by default all type-specific stack functions are "#define"d back to
761 standard stack functions. This results in more streamlined output
762 but retains the type-safety checking possibilities of the original
766 *) The STACK code has been cleaned up, and certain type declarations
767 that didn't make a lot of sense have been brought in line. This has
768 also involved a cleanup of sorts in safestack.h to more correctly
769 map type-safe stack functions onto their plain stack counterparts.
770 This work has also resulted in a variety of "const"ifications of
771 lots of the code, especially "_cmp" operations which should normally
772 be prototyped with "const" parameters anyway.
775 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
776 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
777 (The PRNG state consists of two parts, the large pool 'state' and 'md',
778 where all of 'md' is used each time the PRNG is used, but 'state'
779 is used only indexed by a cyclic counter. As entropy may not be
780 well distributed from the beginning, 'md' is important as a
781 chaining variable. However, the output function chains only half
782 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
783 all of 'md', and seeding with STATE_SIZE dummy bytes will result
784 in all of 'state' being rewritten, with the new values depending
785 on virtually all of 'md'. This overcomes the 80 bit limitation.)
788 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
789 the handshake is continued after ssl_verify_cert_chain();
790 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
791 can lead to 'unexplainable' connection aborts later.
792 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
794 *) Major EVP API cipher revision.
795 Add hooks for extra EVP features. This allows various cipher
796 parameters to be set in the EVP interface. Support added for variable
797 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
798 setting of RC2 and RC5 parameters.
800 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
803 Remove lots of duplicated code from the EVP library. For example *every*
804 cipher init() function handles the 'iv' in the same way according to the
805 cipher mode. They also all do nothing if the 'key' parameter is NULL and
806 for CFB and OFB modes they zero ctx->num.
808 New functionality allows removal of S/MIME code RC2 hack.
810 Most of the routines have the same form and so can be declared in terms
813 By shifting this to the top level EVP_CipherInit() it can be removed from
814 all individual ciphers. If the cipher wants to handle IVs or keys
815 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
818 Change lots of functions like EVP_EncryptUpdate() to now return a
819 value: although software versions of the algorithms cannot fail
820 any installed hardware versions can.
823 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
824 this option is set, tolerate broken clients that send the negotiated
825 protocol version number instead of the requested protocol version
829 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
830 i.e. non-zero for export ciphersuites, zero otherwise.
831 Previous versions had this flag inverted, inconsistent with
832 rsa_tmp_cb (..._TMP_RSA_CB).
833 [Bodo Moeller; problem reported by Amit Chopra]
835 *) Add missing DSA library text string. Work around for some IIS
836 key files with invalid SEQUENCE encoding.
839 *) Add a document (doc/standards.txt) that list all kinds of standards
840 and so on that are implemented in OpenSSL.
843 *) Enhance c_rehash script. Old version would mishandle certificates
844 with the same subject name hash and wouldn't handle CRLs at all.
845 Added -fingerprint option to crl utility, to support new c_rehash
849 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
852 *) Fix for SSL server purpose checking. Server checking was
853 rejecting certificates which had extended key usage present
854 but no ssl client purpose.
855 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
857 *) Make PKCS#12 code work with no password. The PKCS#12 spec
858 is a little unclear about how a blank password is handled.
859 Since the password in encoded as a BMPString with terminating
860 double NULL a zero length password would end up as just the
861 double NULL. However no password at all is different and is
862 handled differently in the PKCS#12 key generation code. NS
863 treats a blank password as zero length. MSIE treats it as no
864 password on export: but it will try both on import. We now do
865 the same: PKCS12_parse() tries zero length and no password if
866 the password is set to "" or NULL (NULL is now a valid password:
867 it wasn't before) as does the pkcs12 application.
870 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
871 perror when PEM_read_bio_X509_REQ fails, the error message must
872 be obtained from the error queue.
875 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
876 it in ERR_remove_state if appropriate, and change ERR_get_state
877 accordingly to avoid race conditions (this is necessary because
878 thread_hash is no longer constant once set).
881 *) Bugfix for linux-elf makefile.one.
884 *) RSA_get_default_method() will now cause a default
885 RSA_METHOD to be chosen if one doesn't exist already.
886 Previously this was only set during a call to RSA_new()
887 or RSA_new_method(NULL) meaning it was possible for
888 RSA_get_default_method() to return NULL.
891 *) Added native name translation to the existing DSO code
892 that will convert (if the flag to do so is set) filenames
893 that are sufficiently small and have no path information
894 into a canonical native form. Eg. "blah" converted to
895 "libblah.so" or "blah.dll" etc.
898 *) New function ERR_error_string_n(e, buf, len) which is like
899 ERR_error_string(e, buf), but writes at most 'len' bytes
900 including the 0 terminator. For ERR_error_string_n, 'buf'
902 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
904 *) CONF library reworked to become more general. A new CONF
905 configuration file reader "class" is implemented as well as a
906 new functions (NCONF_*, for "New CONF") to handle it. The now
907 old CONF_* functions are still there, but are reimplemented to
908 work in terms of the new functions. Also, a set of functions
909 to handle the internal storage of the configuration data is
910 provided to make it easier to write new configuration file
911 reader "classes" (I can definitely see something reading a
912 configuration file in XML format, for example), called _CONF_*,
913 or "the configuration storage API"...
915 The new configuration file reading functions are:
917 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
918 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
920 NCONF_default, NCONF_WIN32
922 NCONF_dump_fp, NCONF_dump_bio
924 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
925 NCONF_new creates a new CONF object. This works in the same way
926 as other interfaces in OpenSSL, like the BIO interface.
927 NCONF_dump_* dump the internal storage of the configuration file,
928 which is useful for debugging. All other functions take the same
929 arguments as the old CONF_* functions wth the exception of the
930 first that must be a `CONF *' instead of a `LHASH *'.
932 To make it easer to use the new classes with the old CONF_* functions,
933 the function CONF_set_default_method is provided.
936 *) Add '-tls1' option to 'openssl ciphers', which was already
937 mentioned in the documentation but had not been implemented.
938 (This option is not yet really useful because even the additional
939 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
942 *) Initial DSO code added into libcrypto for letting OpenSSL (and
943 OpenSSL-based applications) load shared libraries and bind to
944 them in a portable way.
945 [Geoff Thorpe, with contributions from Richard Levitte]
947 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
949 *) Make sure _lrotl and _lrotr are only used with MSVC.
951 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
952 (the default implementation of RAND_status).
954 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
955 to '-clrext' (= clear extensions), as intended and documented.
956 [Bodo Moeller; inconsistency pointed out by Michael Attili
959 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
960 was larger than the MD block size.
961 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
963 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
964 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
965 using the passed key: if the passed key was a private key the result
966 of X509_print(), for example, would be to print out all the private key
970 *) des_quad_cksum() byte order bug fix.
971 [Ulf Möller, using the problem description in krb4-0.9.7, where
972 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
974 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
976 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
978 *) For easily testing in shell scripts whether some command
979 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
980 returns with exit code 0 iff no command of the given name is available.
981 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
982 the output goes to stdout and nothing is printed to stderr.
983 Additional arguments are always ignored.
985 Since for each cipher there is a command of the same name,
986 the 'no-cipher' compilation switches can be tested this way.
988 ('openssl no-XXX' is not able to detect pseudo-commands such
989 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
992 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
995 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
996 is set; it will be thrown away anyway because each handshake creates
998 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
999 to parameters -- in previous versions (since OpenSSL 0.9.3) the
1000 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
1001 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
1004 *) New s_client option -ign_eof: EOF at stdin is ignored, and
1005 'Q' and 'R' lose their special meanings (quit/renegotiate).
1006 This is part of what -quiet does; unlike -quiet, -ign_eof
1007 does not suppress any output.
1010 *) Add compatibility options to the purpose and trust code. The
1011 purpose X509_PURPOSE_ANY is "any purpose" which automatically
1012 accepts a certificate or CA, this was the previous behaviour,
1013 with all the associated security issues.
1015 X509_TRUST_COMPAT is the old trust behaviour: only and
1016 automatically trust self signed roots in certificate store. A
1017 new trust setting X509_TRUST_DEFAULT is used to specify that
1018 a purpose has no associated trust setting and it should instead
1019 use the value in the default purpose.
1022 *) Fix the PKCS#8 DSA private key code so it decodes keys again
1023 and fix a memory leak.
1026 *) In util/mkerr.pl (which implements 'make errors'), preserve
1027 reason strings from the previous version of the .c file, as
1028 the default to have only downcase letters (and digits) in
1029 automatically generated reasons codes is not always appropriate.
1032 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
1033 using strerror. Previously, ERR_reason_error_string() returned
1034 library names as reason strings for SYSerr; but SYSerr is a special
1035 case where small numbers are errno values, not library numbers.
1038 *) Add '-dsaparam' option to 'openssl dhparam' application. This
1039 converts DSA parameters into DH parameters. (When creating parameters,
1040 DSA_generate_parameters is used.)
1043 *) Include 'length' (recommended exponent length) in C code generated
1044 by 'openssl dhparam -C'.
1047 *) The second argument to set_label in perlasm was already being used
1048 so couldn't be used as a "file scope" flag. Moved to third argument
1052 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
1053 instead of RAND_bytes for encryption IVs and salts.
1056 *) Include RAND_status() into RAND_METHOD instead of implementing
1057 it only for md_rand.c Otherwise replacing the PRNG by calling
1058 RAND_set_rand_method would be impossible.
1061 *) Don't let DSA_generate_key() enter an infinite loop if the random
1062 number generation fails.
1065 *) New 'rand' application for creating pseudo-random output.
1068 *) Added configuration support for Linux/IA64
1069 [Rolf Haberrecker <rolf@suse.de>]
1071 *) Assembler module support for Mingw32.
1074 *) Shared library support for HPUX (in shlib/).
1075 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
1077 *) Shared library support for Solaris gcc.
1078 [Lutz Behnke <behnke@trustcenter.de>]
1080 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
1082 *) PKCS7_encrypt() was adding text MIME headers twice because they
1083 were added manually and by SMIME_crlf_copy().
1086 *) In bntest.c don't call BN_rand with zero bits argument.
1087 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
1089 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
1090 case was implemented. This caused BN_div_recp() to fail occasionally.
1093 *) Add an optional second argument to the set_label() in the perl
1094 assembly language builder. If this argument exists and is set
1095 to 1 it signals that the assembler should use a symbol whose
1096 scope is the entire file, not just the current function. This
1097 is needed with MASM which uses the format label:: for this scope.
1098 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
1100 *) Change the ASN1 types so they are typedefs by default. Before
1101 almost all types were #define'd to ASN1_STRING which was causing
1102 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
1106 *) Change names of new functions to the new get1/get0 naming
1107 convention: After 'get1', the caller owns a reference count
1108 and has to call ..._free; 'get0' returns a pointer to some
1109 data structure without incrementing reference counters.
1110 (Some of the existing 'get' functions increment a reference
1111 counter, some don't.)
1112 Similarly, 'set1' and 'add1' functions increase reference
1113 counters or duplicate objects.
1116 *) Allow for the possibility of temp RSA key generation failure:
1117 the code used to assume it always worked and crashed on failure.
1120 *) Fix potential buffer overrun problem in BIO_printf().
1121 [Ulf Möller, using public domain code by Patrick Powell; problem
1122 pointed out by David Sacerdote <das33@cornell.edu>]
1124 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
1125 RAND_egd() and RAND_status(). In the command line application,
1126 the EGD socket can be specified like a seed file using RANDFILE
1130 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
1131 Some CAs (e.g. Verisign) distribute certificates in this form.
1134 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
1135 list to exclude them. This means that no special compilation option
1136 is needed to use anonymous DH: it just needs to be included in the
1140 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
1141 EVP_MD_type. The old functionality is available in a new macro called
1142 EVP_MD_md(). Change code that uses it and update docs.
1145 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
1146 where the 'void *' argument is replaced by a function pointer argument.
1147 Previously 'void *' was abused to point to functions, which works on
1148 many platforms, but is not correct. As these functions are usually
1149 called by macros defined in OpenSSL header files, most source code
1150 should work without changes.
1153 *) <openssl/opensslconf.h> (which is created by Configure) now contains
1154 sections with information on -D... compiler switches used for
1155 compiling the library so that applications can see them. To enable
1156 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
1157 must be defined. E.g.,
1158 #define OPENSSL_ALGORITHM_DEFINES
1159 #include <openssl/opensslconf.h>
1160 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
1161 [Richard Levitte, Ulf and Bodo Möller]
1163 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
1167 *) Change the 'other' type in certificate aux info to a STACK_OF
1168 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
1169 the required ASN1 format: arbitrary types determined by an OID.
1172 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
1173 argument to 'req'. This is not because the function is newer or
1174 better than others it just uses the work 'NEW' in the certificate
1175 request header lines. Some software needs this.
1178 *) Reorganise password command line arguments: now passwords can be
1179 obtained from various sources. Delete the PEM_cb function and make
1180 it the default behaviour: i.e. if the callback is NULL and the
1181 usrdata argument is not NULL interpret it as a null terminated pass
1182 phrase. If usrdata and the callback are NULL then the pass phrase
1183 is prompted for as usual.
1186 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
1187 the support is automatically enabled. The resulting binaries will
1188 autodetect the card and use it if present.
1189 [Ben Laurie and Compaq Inc.]
1191 *) Work around for Netscape hang bug. This sends certificate request
1192 and server done in one record. Since this is perfectly legal in the
1193 SSL/TLS protocol it isn't a "bug" option and is on by default. See
1194 the bugs/SSLv3 entry for more info.
1197 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
1200 *) Add -rand argument to smime and pkcs12 applications and read/write
1204 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
1207 *) Add command line password options to the remaining applications.
1210 *) Bug fix for BN_div_recp() for numerators with an even number of
1214 *) More tests in bntest.c, and changed test_bn output.
1217 *) ./config recognizes MacOS X now.
1220 *) Bug fix for BN_div() when the first words of num and divsor are
1221 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
1224 *) Add support for various broken PKCS#8 formats, and command line
1225 options to produce them.
1228 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
1229 get temporary BIGNUMs from a BN_CTX.
1232 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
1236 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
1237 include a #define from the old name to the new. The original intent
1238 was that statically linked binaries could for example just call
1239 SSLeay_add_all_ciphers() to just add ciphers to the table and not
1240 link with digests. This never worked becayse SSLeay_add_all_digests()
1241 and SSLeay_add_all_ciphers() were in the same source file so calling
1242 one would link with the other. They are now in separate source files.
1245 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
1248 *) Use a less unusual form of the Miller-Rabin primality test (it used
1249 a binary algorithm for exponentiation integrated into the Miller-Rabin
1250 loop, our standard modexp algorithms are faster).
1253 *) Support for the EBCDIC character set completed.
1254 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
1256 *) Source code cleanups: use const where appropriate, eliminate casts,
1257 use void * instead of char * in lhash.
1260 *) Bugfix: ssl3_send_server_key_exchange was not restartable
1261 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
1262 this the server could overwrite ephemeral keys that the client
1266 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
1267 using 50 iterations of the Rabin-Miller test.
1269 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
1270 iterations of the Rabin-Miller test as required by the appendix
1271 to FIPS PUB 186[-1]) instead of DSA_is_prime.
1272 As BN_is_prime_fasttest includes trial division, DSA parameter
1273 generation becomes much faster.
1275 This implies a change for the callback functions in DSA_is_prime
1276 and DSA_generate_parameters: The callback function is called once
1277 for each positive witness in the Rabin-Miller test, not just
1278 occasionally in the inner loop; and the parameters to the
1279 callback function now provide an iteration count for the outer
1280 loop rather than for the current invocation of the inner loop.
1281 DSA_generate_parameters additionally can call the callback
1282 function with an 'iteration count' of -1, meaning that a
1283 candidate has passed the trial division test (when q is generated
1284 from an application-provided seed, trial division is skipped).
1287 *) New function BN_is_prime_fasttest that optionally does trial
1288 division before starting the Rabin-Miller test and has
1289 an additional BN_CTX * argument (whereas BN_is_prime always
1290 has to allocate at least one BN_CTX).
1291 'callback(1, -1, cb_arg)' is called when a number has passed the
1292 trial division stage.
1295 *) Fix for bug in CRL encoding. The validity dates weren't being handled
1299 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
1302 *) New function BN_pseudo_rand().
1305 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
1306 bignum version of BN_from_montgomery() with the working code from
1307 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
1311 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
1312 made it impossible to use the same SSL_SESSION data structure in
1313 SSL2 clients in multiple threads.
1316 *) The return value of RAND_load_file() no longer counts bytes obtained
1317 by stat(). RAND_load_file(..., -1) is new and uses the complete file
1318 to seed the PRNG (previously an explicit byte count was required).
1319 [Ulf Möller, Bodo Möller]
1321 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
1322 used (char *) instead of (void *) and had casts all over the place.
1325 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
1328 *) Retain source code compatibility for BN_prime_checks macro:
1329 BN_is_prime(..., BN_prime_checks, ...) now uses
1330 BN_prime_checks_for_size to determine the appropriate number of
1331 Rabin-Miller iterations.
1334 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
1335 DH_CHECK_P_NOT_SAFE_PRIME.
1336 (Check if this is true? OpenPGP calls them "strong".)
1339 *) Merge the functionality of "dh" and "gendh" programs into a new program
1340 "dhparam". The old programs are retained for now but will handle DH keys
1341 (instead of parameters) in future.
1344 *) Make the ciphers, s_server and s_client programs check the return values
1345 when a new cipher list is set.
1348 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
1349 ciphers. Before when the 56bit ciphers were enabled the sorting was
1352 The syntax for the cipher sorting has been extended to support sorting by
1353 cipher-strength (using the strength_bits hard coded in the tables).
1354 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
1356 Fix a bug in the cipher-command parser: when supplying a cipher command
1357 string with an "undefined" symbol (neither command nor alphanumeric
1358 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
1359 an error is flagged.
1361 Due to the strength-sorting extension, the code of the
1362 ssl_create_cipher_list() function was completely rearranged. I hope that
1363 the readability was also increased :-)
1364 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
1366 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
1367 for the first serial number and places 2 in the serial number file. This
1368 avoids problems when the root CA is created with serial number zero and
1369 the first user certificate has the same issuer name and serial number
1373 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
1374 the new code. Add documentation for this stuff.
1377 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
1378 X509_*() to X509at_*() on the grounds that they don't handle X509
1379 structures and behave in an analagous way to the X509v3 functions:
1380 they shouldn't be called directly but wrapper functions should be used
1383 So we also now have some wrapper functions that call the X509at functions
1384 when passed certificate requests. (TO DO: similar things can be done with
1385 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
1386 things. Some of these need some d2i or i2d and print functionality
1387 because they handle more complex structures.)
1390 *) Add missing #ifndefs that caused missing symbols when building libssl
1391 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
1392 NO_RSA in ssl/s2*.c.
1393 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
1395 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
1396 has a return value which indicates the quality of the random data
1397 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
1398 error queue. New function RAND_pseudo_bytes() generates output that is
1399 guaranteed to be unique but not unpredictable. RAND_add is like
1400 RAND_seed, but takes an extra argument for an entropy estimate
1401 (RAND_seed always assumes full entropy).
1404 *) Do more iterations of Rabin-Miller probable prime test (specifically,
1405 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
1406 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
1407 in crypto/bn/bn_prime.c for the complete table). This guarantees a
1408 false-positive rate of at most 2^-80 for random input.
1411 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
1414 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
1415 in the 0.9.5 release), this returns the chain
1416 from an X509_CTX structure with a dup of the stack and all
1417 the X509 reference counts upped: so the stack will exist
1418 after X509_CTX_cleanup() has been called. Modify pkcs12.c
1421 Also make SSL_SESSION_print() print out the verify return
1425 *) Add manpage for the pkcs12 command. Also change the default
1426 behaviour so MAC iteration counts are used unless the new
1427 -nomaciter option is used. This improves file security and
1428 only older versions of MSIE (4.0 for example) need it.
1431 *) Honor the no-xxx Configure options when creating .DEF files.
1434 *) Add PKCS#10 attributes to field table: challengePassword,
1435 unstructuredName and unstructuredAddress. These are taken from
1436 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
1437 international characters are used.
1439 More changes to X509_ATTRIBUTE code: allow the setting of types
1440 based on strings. Remove the 'loc' parameter when adding
1441 attributes because these will be a SET OF encoding which is sorted
1445 *) Initial changes to the 'req' utility to allow request generation
1446 automation. This will allow an application to just generate a template
1447 file containing all the field values and have req construct the
1450 Initial support for X509_ATTRIBUTE handling. Stacks of these are
1451 used all over the place including certificate requests and PKCS#7
1452 structures. They are currently handled manually where necessary with
1453 some primitive wrappers for PKCS#7. The new functions behave in a
1454 manner analogous to the X509 extension functions: they allow
1455 attributes to be looked up by NID and added.
1457 Later something similar to the X509V3 code would be desirable to
1458 automatically handle the encoding, decoding and printing of the
1459 more complex types. The string types like challengePassword can
1460 be handled by the string table functions.
1462 Also modified the multi byte string table handling. Now there is
1463 a 'global mask' which masks out certain types. The table itself
1464 can use the flag STABLE_NO_MASK to ignore the mask setting: this
1465 is useful when for example there is only one permissible type
1466 (as in countryName) and using the mask might result in no valid
1470 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
1471 SSL_get_peer_finished to allow applications to obtain the latest
1472 Finished messages sent to the peer or expected from the peer,
1473 respectively. (SSL_get_peer_finished is usually the Finished message
1474 actually received from the peer, otherwise the protocol will be aborted.)
1476 As the Finished message are message digests of the complete handshake
1477 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
1478 be used for external authentication procedures when the authentication
1479 provided by SSL/TLS is not desired or is not enough.
1482 *) Enhanced support for Alpha Linux is added. Now ./config checks if
1483 the host supports BWX extension and if Compaq C is present on the
1484 $PATH. Just exploiting of the BWX extension results in 20-30%
1485 performance kick for some algorithms, e.g. DES and RC4 to mention
1486 a couple. Compaq C in turn generates ~20% faster code for MD5 and
1490 *) Add support for MS "fast SGC". This is arguably a violation of the
1491 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
1492 weak crypto and after checking the certificate is SGC a second one
1493 with strong crypto. MS SGC stops the first handshake after receiving
1494 the server certificate message and sends a second client hello. Since
1495 a server will typically do all the time consuming operations before
1496 expecting any further messages from the client (server key exchange
1497 is the most expensive) there is little difference between the two.
1499 To get OpenSSL to support MS SGC we have to permit a second client
1500 hello message after we have sent server done. In addition we have to
1501 reset the MAC if we do get this second client hello.
1504 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
1505 if a DER encoded private key is RSA or DSA traditional format. Changed
1506 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
1507 format DER encoded private key. Newer code should use PKCS#8 format which
1508 has the key type encoded in the ASN1 structure. Added DER private key
1509 support to pkcs8 application.
1512 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
1513 ciphersuites has been selected (as required by the SSL 3/TLS 1
1514 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
1515 is set, we interpret this as a request to violate the specification
1516 (the worst that can happen is a handshake failure, and 'correct'
1517 behaviour would result in a handshake failure anyway).
1520 *) In SSL_CTX_add_session, take into account that there might be multiple
1521 SSL_SESSION structures with the same session ID (e.g. when two threads
1522 concurrently obtain them from an external cache).
1523 The internal cache can handle only one SSL_SESSION with a given ID,
1524 so if there's a conflict, we now throw out the old one to achieve
1528 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
1529 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
1530 some routines that use cipher OIDs: some ciphers do not have OIDs
1531 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
1535 *) Simplify the trust setting structure and code. Now we just have
1536 two sequences of OIDs for trusted and rejected settings. These will
1537 typically have values the same as the extended key usage extension
1538 and any application specific purposes.
1540 The trust checking code now has a default behaviour: it will just
1541 check for an object with the same NID as the passed id. Functions can
1542 be provided to override either the default behaviour or the behaviour
1543 for a given id. SSL client, server and email already have functions
1544 in place for compatibility: they check the NID and also return "trusted"
1545 if the certificate is self signed.
1548 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
1549 traditional format into an EVP_PKEY structure.
1552 *) Add a password callback function PEM_cb() which either prompts for
1553 a password if usr_data is NULL or otherwise assumes it is a null
1554 terminated password. Allow passwords to be passed on command line
1555 environment or config files in a few more utilities.
1558 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
1559 keys. Add some short names for PKCS#8 PBE algorithms and allow them
1560 to be specified on the command line for the pkcs8 and pkcs12 utilities.
1561 Update documentation.
1564 *) Support for ASN1 "NULL" type. This could be handled before by using
1565 ASN1_TYPE but there wasn't any function that would try to read a NULL
1566 and produce an error if it couldn't. For compatibility we also have
1567 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
1568 don't allocate anything because they don't need to.
1571 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
1573 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
1575 *) Rebuild of the memory allocation routines used by OpenSSL code and
1576 possibly others as well. The purpose is to make an interface that
1577 provide hooks so anyone can build a separate set of allocation and
1578 deallocation routines to be used by OpenSSL, for example memory
1579 pool implementations, or something else, which was previously hard
1580 since Malloc(), Realloc() and Free() were defined as macros having
1581 the values malloc, realloc and free, respectively (except for Win32
1582 compilations). The same is provided for memory debugging code.
1583 OpenSSL already comes with functionality to find memory leaks, but
1584 this gives people a chance to debug other memory problems.
1586 With these changes, a new set of functions and macros have appeared:
1588 CRYPTO_set_mem_debug_functions() [F]
1589 CRYPTO_get_mem_debug_functions() [F]
1590 CRYPTO_dbg_set_options() [F]
1591 CRYPTO_dbg_get_options() [F]
1592 CRYPTO_malloc_debug_init() [M]
1594 The memory debug functions are NULL by default, unless the library
1595 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
1596 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
1597 gives the standard debugging functions that come with OpenSSL) or
1598 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
1599 provided by the library user) must be used. When the standard
1600 debugging functions are used, CRYPTO_dbg_set_options can be used to
1601 request additional information:
1602 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
1603 the CRYPTO_MDEBUG_xxx macro when compiling the library.
1605 Also, things like CRYPTO_set_mem_functions will always give the
1606 expected result (the new set of functions is used for allocation
1607 and deallocation) at all times, regardless of platform and compiler
1610 To finish it up, some functions that were never use in any other
1611 way than through macros have a new API and new semantic:
1614 CRYPTO_dbg_realloc()
1617 All macros of value have retained their old syntax.
1618 [Richard Levitte and Bodo Moeller]
1620 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
1621 ordering of SMIMECapabilities wasn't in "strength order" and there
1622 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
1626 *) Some ASN1 types with illegal zero length encoding (INTEGER,
1627 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
1628 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
1630 *) Merge in my S/MIME library for OpenSSL. This provides a simple
1631 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
1632 functionality to handle multipart/signed properly) and a utility
1633 called 'smime' to call all this stuff. This is based on code I
1634 originally wrote for Celo who have kindly allowed it to be
1635 included in OpenSSL.
1638 *) Add variants des_set_key_checked and des_set_key_unchecked of
1639 des_set_key (aka des_key_sched). Global variable des_check_key
1640 decides which of these is called by des_set_key; this way
1641 des_check_key behaves as it always did, but applications and
1642 the library itself, which was buggy for des_check_key == 1,
1643 have a cleaner way to pick the version they need.
1646 *) New function PKCS12_newpass() which changes the password of a
1650 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
1651 dynamic mix. In both cases the ids can be used as an index into the
1652 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
1653 functions so they accept a list of the field values and the
1654 application doesn't need to directly manipulate the X509_TRUST
1658 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
1662 *) Modify the way the V3 extension code looks up extensions. This now
1663 works in a similar way to the object code: we have some "standard"
1664 extensions in a static table which is searched with OBJ_bsearch()
1665 and the application can add dynamic ones if needed. The file
1666 crypto/x509v3/ext_dat.h now has the info: this file needs to be
1667 updated whenever a new extension is added to the core code and kept
1668 in ext_nid order. There is a simple program 'tabtest.c' which checks
1669 this. New extensions are not added too often so this file can readily
1670 be maintained manually.
1672 There are two big advantages in doing things this way. The extensions
1673 can be looked up immediately and no longer need to be "added" using
1674 X509V3_add_standard_extensions(): this function now does nothing.
1675 [Side note: I get *lots* of email saying the extension code doesn't
1676 work because people forget to call this function]
1677 Also no dynamic allocation is done unless new extensions are added:
1678 so if we don't add custom extensions there is no need to call
1679 X509V3_EXT_cleanup().
1682 *) Modify enc utility's salting as follows: make salting the default. Add a
1683 magic header, so unsalted files fail gracefully instead of just decrypting
1684 to garbage. This is because not salting is a big security hole, so people
1685 should be discouraged from doing it.
1688 *) Fixes and enhancements to the 'x509' utility. It allowed a message
1689 digest to be passed on the command line but it only used this
1690 parameter when signing a certificate. Modified so all relevant
1691 operations are affected by the digest parameter including the
1692 -fingerprint and -x509toreq options. Also -x509toreq choked if a
1693 DSA key was used because it didn't fix the digest.
1696 *) Initial certificate chain verify code. Currently tests the untrusted
1697 certificates for consistency with the verify purpose (which is set
1698 when the X509_STORE_CTX structure is set up) and checks the pathlength.
1700 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
1701 this is because it will reject chains with invalid extensions whereas
1702 every previous version of OpenSSL and SSLeay made no checks at all.
1704 Trust code: checks the root CA for the relevant trust settings. Trust
1705 settings have an initial value consistent with the verify purpose: e.g.
1706 if the verify purpose is for SSL client use it expects the CA to be
1707 trusted for SSL client use. However the default value can be changed to
1708 permit custom trust settings: one example of this would be to only trust
1709 certificates from a specific "secure" set of CAs.
1711 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
1712 which should be used for version portability: especially since the
1713 verify structure is likely to change more often now.
1715 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
1716 to set them. If not set then assume SSL clients will verify SSL servers
1719 Two new options to the verify program: -untrusted allows a set of
1720 untrusted certificates to be passed in and -purpose which sets the
1721 intended purpose of the certificate. If a purpose is set then the
1722 new chain verify code is used to check extension consistency.
1725 *) Support for the authority information access extension.
1728 *) Modify RSA and DSA PEM read routines to transparently handle
1729 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
1730 public keys in a format compatible with certificate
1731 SubjectPublicKeyInfo structures. Unfortunately there were already
1732 functions called *_PublicKey_* which used various odd formats so
1733 these are retained for compatibility: however the DSA variants were
1734 never in a public release so they have been deleted. Changed dsa/rsa
1735 utilities to handle the new format: note no releases ever handled public
1736 keys so we should be OK.
1738 The primary motivation for this change is to avoid the same fiasco
1739 that dogs private keys: there are several incompatible private key
1740 formats some of which are standard and some OpenSSL specific and
1741 require various evil hacks to allow partial transparent handling and
1742 even then it doesn't work with DER formats. Given the option anything
1743 other than PKCS#8 should be dumped: but the other formats have to
1744 stay in the name of compatibility.
1746 With public keys and the benefit of hindsight one standard format
1747 is used which works with EVP_PKEY, RSA or DSA structures: though
1748 it clearly returns an error if you try to read the wrong kind of key.
1750 Added a -pubkey option to the 'x509' utility to output the public key.
1751 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
1752 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
1753 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
1754 that do the same as the EVP_PKEY_assign_*() except they up the
1755 reference count of the added key (they don't "swallow" the
1759 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
1760 CRLs would fail if the file contained no certificates or no CRLs:
1761 added a new function to read in both types and return the number
1762 read: this means that if none are read it will be an error. The
1763 DER versions of the certificate and CRL reader would always fail
1764 because it isn't possible to mix certificates and CRLs in DER format
1765 without choking one or the other routine. Changed this to just read
1766 a certificate: this is the best we can do. Also modified the code
1767 in apps/verify.c to take notice of return codes: it was previously
1768 attempting to read in certificates from NULL pointers and ignoring
1769 any errors: this is one reason why the cert and CRL reader seemed
1770 to work. It doesn't check return codes from the default certificate
1771 routines: these may well fail if the certificates aren't installed.
1774 *) Code to support otherName option in GeneralName.
1777 *) First update to verify code. Change the verify utility
1778 so it warns if it is passed a self signed certificate:
1779 for consistency with the normal behaviour. X509_verify
1780 has been modified to it will now verify a self signed
1781 certificate if *exactly* the same certificate appears
1782 in the store: it was previously impossible to trust a
1783 single self signed certificate. This means that:
1784 openssl verify ss.pem
1785 now gives a warning about a self signed certificate but
1786 openssl verify -CAfile ss.pem ss.pem
1790 *) For servers, store verify_result in SSL_SESSION data structure
1791 (and add it to external session representation).
1792 This is needed when client certificate verifications fails,
1793 but an application-provided verification callback (set by
1794 SSL_CTX_set_cert_verify_callback) allows accepting the session
1795 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
1796 but returns 1): When the session is reused, we have to set
1797 ssl->verify_result to the appropriate error code to avoid
1799 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
1801 *) Fix a bug in the new PKCS#7 code: it didn't consider the
1802 case in PKCS7_dataInit() where the signed PKCS7 structure
1803 didn't contain any existing data because it was being created.
1804 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
1806 *) Add a salt to the key derivation routines in enc.c. This
1807 forms the first 8 bytes of the encrypted file. Also add a
1808 -S option to allow a salt to be input on the command line.
1811 *) New function X509_cmp(). Oddly enough there wasn't a function
1812 to compare two certificates. We do this by working out the SHA1
1813 hash and comparing that. X509_cmp() will be needed by the trust
1817 *) SSL_get1_session() is like SSL_get_session(), but increments
1818 the reference count in the SSL_SESSION returned.
1819 [Geoff Thorpe <geoff@eu.c2.net>]
1821 *) Fix for 'req': it was adding a null to request attributes.
1822 Also change the X509_LOOKUP and X509_INFO code to handle
1823 certificate auxiliary information.
1826 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
1830 *) Add the possibility to add extra information to the memory leak
1831 detecting output, to form tracebacks, showing from where each
1832 allocation was originated: CRYPTO_push_info("constant string") adds
1833 the string plus current file name and line number to a per-thread
1834 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
1835 is like calling CYRPTO_pop_info() until the stack is empty.
1836 Also updated memory leak detection code to be multi-thread-safe.
1839 *) Add options -text and -noout to pkcs7 utility and delete the
1840 encryption options which never did anything. Update docs.
1843 *) Add options to some of the utilities to allow the pass phrase
1844 to be included on either the command line (not recommended on
1845 OSes like Unix) or read from the environment. Update the
1846 manpages and fix a few bugs.
1849 *) Add a few manpages for some of the openssl commands.
1852 *) Fix the -revoke option in ca. It was freeing up memory twice,
1853 leaking and not finding already revoked certificates.
1856 *) Extensive changes to support certificate auxiliary information.
1857 This involves the use of X509_CERT_AUX structure and X509_AUX
1858 functions. An X509_AUX function such as PEM_read_X509_AUX()
1859 can still read in a certificate file in the usual way but it
1860 will also read in any additional "auxiliary information". By
1861 doing things this way a fair degree of compatibility can be
1862 retained: existing certificates can have this information added
1863 using the new 'x509' options.
1865 Current auxiliary information includes an "alias" and some trust
1866 settings. The trust settings will ultimately be used in enhanced
1867 certificate chain verification routines: currently a certificate
1868 can only be trusted if it is self signed and then it is trusted
1872 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
1873 The problem was that one of the replacement routines had not been working
1874 since SSLeay releases. For now the offending routine has been replaced
1875 with non-optimised assembler. Even so, this now gives around 95%
1876 performance improvement for 1024 bit RSA signs.
1879 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
1880 handling. Most clients have the effective key size in bits equal to
1881 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
1882 A few however don't do this and instead use the size of the decrypted key
1883 to determine the RC2 key length and the AlgorithmIdentifier to determine
1884 the effective key length. In this case the effective key length can still
1885 be 40 bits but the key length can be 168 bits for example. This is fixed
1886 by manually forcing an RC2 key into the EVP_PKEY structure because the
1887 EVP code can't currently handle unusual RC2 key sizes: it always assumes
1888 the key length and effective key length are equal.
1891 *) Add a bunch of functions that should simplify the creation of
1892 X509_NAME structures. Now you should be able to do:
1893 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
1894 and have it automatically work out the correct field type and fill in
1895 the structures. The more adventurous can try:
1896 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
1897 and it will (hopefully) work out the correct multibyte encoding.
1900 *) Change the 'req' utility to use the new field handling and multibyte
1901 copy routines. Before the DN field creation was handled in an ad hoc
1902 way in req, ca, and x509 which was rather broken and didn't support
1903 BMPStrings or UTF8Strings. Since some software doesn't implement
1904 BMPStrings or UTF8Strings yet, they can be enabled using the config file
1905 using the dirstring_type option. See the new comment in the default
1906 openssl.cnf for more info.
1909 *) Make crypto/rand/md_rand.c more robust:
1910 - Assure unique random numbers after fork().
1911 - Make sure that concurrent threads access the global counter and
1912 md serializably so that we never lose entropy in them
1913 or use exactly the same state in multiple threads.
1914 Access to the large state is not always serializable because
1915 the additional locking could be a performance killer, and
1916 md should be large enough anyway.
1919 *) New file apps/app_rand.c with commonly needed functionality
1920 for handling the random seed file.
1922 Use the random seed file in some applications that previously did not:
1924 dsaparam -genkey (which also ignored its '-rand' option),
1927 x509 (when signing).
1928 Except on systems with /dev/urandom, it is crucial to have a random
1929 seed file at least for key creation, DSA signing, and for DH exchanges;
1930 for RSA signatures we could do without one.
1932 gendh and gendsa (unlike genrsa) used to read only the first byte
1933 of each file listed in the '-rand' option. The function as previously
1934 found in genrsa is now in app_rand.c and is used by all programs
1935 that support '-rand'.
1938 *) In RAND_write_file, use mode 0600 for creating files;
1939 don't just chmod when it may be too late.
1942 *) Report an error from X509_STORE_load_locations
1943 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
1946 *) New function ASN1_mbstring_copy() this copies a string in either
1947 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
1948 into an ASN1_STRING type. A mask of permissible types is passed
1949 and it chooses the "minimal" type to use or an error if not type
1953 *) Add function equivalents to the various macros in asn1.h. The old
1954 macros are retained with an M_ prefix. Code inside the library can
1955 use the M_ macros. External code (including the openssl utility)
1956 should *NOT* in order to be "shared library friendly".
1959 *) Add various functions that can check a certificate's extensions
1960 to see if it usable for various purposes such as SSL client,
1961 server or S/MIME and CAs of these types. This is currently
1962 VERY EXPERIMENTAL but will ultimately be used for certificate chain
1963 verification. Also added a -purpose flag to x509 utility to
1964 print out all the purposes.
1967 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
1971 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
1972 for, obtain and decode and extension and obtain its critical flag.
1973 This allows all the necessary extension code to be handled in a
1974 single function call.
1977 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
1978 platforms. See crypto/rc4/rc4_enc.c for further details.
1981 *) New -noout option to asn1parse. This causes no output to be produced
1982 its main use is when combined with -strparse and -out to extract data
1983 from a file (which may not be in ASN.1 format).
1986 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
1987 when producing the local key id.
1988 [Richard Levitte <levitte@stacken.kth.se>]
1990 *) New option -dhparam in s_server. This allows a DH parameter file to be
1991 stated explicitly. If it is not stated then it tries the first server
1992 certificate file. The previous behaviour hard coded the filename
1996 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
1997 a public key to be input or output. For example:
1998 openssl rsa -in key.pem -pubout -out pubkey.pem
1999 Also added necessary DSA public key functions to handle this.
2002 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
2003 in the message. This was handled by allowing
2004 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
2005 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
2007 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
2008 to the end of the strings whereas this didn't. This would cause problems
2009 if strings read with d2i_ASN1_bytes() were later modified.
2010 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
2012 *) Fix for base64 decode bug. When a base64 bio reads only one line of
2013 data and it contains EOF it will end up returning an error. This is
2014 caused by input 46 bytes long. The cause is due to the way base64
2015 BIOs find the start of base64 encoded data. They do this by trying a
2016 trial decode on each line until they find one that works. When they
2017 do a flag is set and it starts again knowing it can pass all the
2018 data directly through the decoder. Unfortunately it doesn't reset
2019 the context it uses. This means that if EOF is reached an attempt
2020 is made to pass two EOFs through the context and this causes the
2021 resulting error. This can also cause other problems as well. As is
2022 usual with these problems it takes *ages* to find and the fix is
2023 trivial: move one line.
2024 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
2026 *) Ugly workaround to get s_client and s_server working under Windows. The
2027 old code wouldn't work because it needed to select() on sockets and the
2028 tty (for keypresses and to see if data could be written). Win32 only
2029 supports select() on sockets so we select() with a 1s timeout on the
2030 sockets and then see if any characters are waiting to be read, if none
2031 are present then we retry, we also assume we can always write data to
2032 the tty. This isn't nice because the code then blocks until we've
2033 received a complete line of data and it is effectively polling the
2034 keyboard at 1s intervals: however it's quite a bit better than not
2035 working at all :-) A dedicated Windows application might handle this
2036 with an event loop for example.
2039 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
2040 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
2041 will be called when RSA_sign() and RSA_verify() are used. This is useful
2042 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
2043 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
2044 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
2045 This necessitated the support of an extra signature type NID_md5_sha1
2046 for SSL signatures and modifications to the SSL library to use it instead
2047 of calling RSA_public_decrypt() and RSA_private_encrypt().
2050 *) Add new -verify -CAfile and -CApath options to the crl program, these
2051 will lookup a CRL issuers certificate and verify the signature in a
2052 similar way to the verify program. Tidy up the crl program so it
2053 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
2054 less strict. It will now permit CRL extensions even if it is not
2055 a V2 CRL: this will allow it to tolerate some broken CRLs.
2058 *) Initialize all non-automatic variables each time one of the openssl
2059 sub-programs is started (this is necessary as they may be started
2060 multiple times from the "OpenSSL>" prompt).
2061 [Lennart Bang, Bodo Moeller]
2063 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
2064 removing all other RSA functionality (this is what NO_RSA does). This
2065 is so (for example) those in the US can disable those operations covered
2066 by the RSA patent while allowing storage and parsing of RSA keys and RSA
2070 *) Non-copying interface to BIO pairs.
2071 (still largely untested)
2074 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
2075 ASCII string. This was handled independently in various places before.
2078 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
2079 UTF8 strings a character at a time.
2082 *) Use client_version from client hello to select the protocol
2083 (s23_srvr.c) and for RSA client key exchange verification
2084 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
2087 *) Add various utility functions to handle SPKACs, these were previously
2088 handled by poking round in the structure internals. Added new function
2089 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
2090 print, verify and generate SPKACs. Based on an original idea from
2091 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
2094 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
2097 *) Allow the config file extension section to be overwritten on the
2098 command line. Based on an original idea from Massimiliano Pala
2099 <madwolf@comune.modena.it>. The new option is called -extensions
2100 and can be applied to ca, req and x509. Also -reqexts to override
2101 the request extensions in req and -crlexts to override the crl extensions
2105 *) Add new feature to the SPKAC handling in ca. Now you can include
2106 the same field multiple times by preceding it by "XXXX." for example:
2109 this is the same syntax as used in the req config file.
2112 *) Allow certificate extensions to be added to certificate requests. These
2113 are specified in a 'req_extensions' option of the req section of the
2114 config file. They can be printed out with the -text option to req but
2115 are otherwise ignored at present.
2118 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
2119 data read consists of only the final block it would not decrypted because
2120 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
2121 A misplaced 'break' also meant the decrypted final block might not be
2122 copied until the next read.
2125 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
2126 a few extra parameters to the DH structure: these will be useful if
2127 for example we want the value of 'q' or implement X9.42 DH.
2130 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
2131 provides hooks that allow the default DSA functions or functions on a
2132 "per key" basis to be replaced. This allows hardware acceleration and
2133 hardware key storage to be handled without major modification to the
2134 library. Also added low level modexp hooks and CRYPTO_EX structure and
2135 associated functions.
2138 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
2139 as "read only": it can't be written to and the buffer it points to will
2140 not be freed. Reading from a read only BIO is much more efficient than
2141 a normal memory BIO. This was added because there are several times when
2142 an area of memory needs to be read from a BIO. The previous method was
2143 to create a memory BIO and write the data to it, this results in two
2144 copies of the data and an O(n^2) reading algorithm. There is a new
2145 function BIO_new_mem_buf() which creates a read only memory BIO from
2146 an area of memory. Also modified the PKCS#7 routines to use read only
2150 *) Bugfix: ssl23_get_client_hello did not work properly when called in
2151 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
2152 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
2153 but a retry condition occured while trying to read the rest.
2156 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
2157 NID_pkcs7_encrypted by default: this was wrong since this should almost
2158 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
2159 the encrypted data type: this is a more sensible place to put it and it
2160 allows the PKCS#12 code to be tidied up that duplicated this
2164 *) Changed obj_dat.pl script so it takes its input and output files on
2165 the command line. This should avoid shell escape redirection problems
2169 *) Initial support for certificate extension requests, these are included
2170 in things like Xenroll certificate requests. Included functions to allow
2171 extensions to be obtained and added.
2174 *) -crlf option to s_client and s_server for sending newlines as
2175 CRLF (as required by many protocols).
2178 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
2180 *) Install libRSAglue.a when OpenSSL is built with RSAref.
2181 [Ralf S. Engelschall]
2183 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
2184 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
2186 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
2190 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
2191 DH parameters/keys (q is lost during that conversion, but the resulting
2192 DH parameters contain its length).
2194 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
2195 much faster than DH_generate_parameters (which creates parameters
2196 where p = 2*q + 1), and also the smaller q makes DH computations
2197 much more efficient (160-bit exponentiation instead of 1024-bit
2198 exponentiation); so this provides a convenient way to support DHE
2199 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
2200 utter importance to use
2201 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
2203 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
2204 when such DH parameters are used, because otherwise small subgroup
2205 attacks may become possible!
2208 *) Avoid memory leak in i2d_DHparams.
2211 *) Allow the -k option to be used more than once in the enc program:
2212 this allows the same encrypted message to be read by multiple recipients.
2215 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
2216 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
2217 it will always use the numerical form of the OID, even if it has a short
2221 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
2222 method only got called if p,q,dmp1,dmq1,iqmp components were present,
2223 otherwise bn_mod_exp was called. In the case of hardware keys for example
2224 no private key components need be present and it might store extra data
2225 in the RSA structure, which cannot be accessed from bn_mod_exp.
2226 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
2227 private key operations.
2230 *) Added support for SPARC Linux.
2233 *) pem_password_cb function type incompatibly changed from
2234 typedef int pem_password_cb(char *buf, int size, int rwflag);
2236 ....(char *buf, int size, int rwflag, void *userdata);
2237 so that applications can pass data to their callbacks:
2238 The PEM[_ASN1]_{read,write}... functions and macros now take an
2239 additional void * argument, which is just handed through whenever
2240 the password callback is called.
2241 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
2243 New function SSL_CTX_set_default_passwd_cb_userdata.
2245 Compatibility note: As many C implementations push function arguments
2246 onto the stack in reverse order, the new library version is likely to
2247 interoperate with programs that have been compiled with the old
2248 pem_password_cb definition (PEM_whatever takes some data that
2249 happens to be on the stack as its last argument, and the callback
2250 just ignores this garbage); but there is no guarantee whatsoever that
2253 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
2254 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
2255 problems not only on Windows, but also on some Unix platforms.
2256 To avoid problematic command lines, these definitions are now in an
2257 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
2258 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
2261 *) MIPS III/IV assembler module is reimplemented.
2264 *) More DES library cleanups: remove references to srand/rand and
2265 delete an unused file.
2268 *) Add support for the the free Netwide assembler (NASM) under Win32,
2269 since not many people have MASM (ml) and it can be hard to obtain.
2270 This is currently experimental but it seems to work OK and pass all
2271 the tests. Check out INSTALL.W32 for info.
2274 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
2275 without temporary keys kept an extra copy of the server key,
2276 and connections with temporary keys did not free everything in case
2280 *) New function RSA_check_key and new openssl rsa option -check
2281 for verifying the consistency of RSA keys.
2282 [Ulf Moeller, Bodo Moeller]
2284 *) Various changes to make Win32 compile work:
2285 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
2286 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
2287 comparison" warnings.
2288 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
2291 *) Add a debugging option to PKCS#5 v2 key generation function: when
2292 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
2293 derived keys are printed to stderr.
2296 *) Copy the flags in ASN1_STRING_dup().
2297 [Roman E. Pavlov <pre@mo.msk.ru>]
2299 *) The x509 application mishandled signing requests containing DSA
2300 keys when the signing key was also DSA and the parameters didn't match.
2302 It was supposed to omit the parameters when they matched the signing key:
2303 the verifying software was then supposed to automatically use the CA's
2304 parameters if they were absent from the end user certificate.
2306 Omitting parameters is no longer recommended. The test was also
2307 the wrong way round! This was probably due to unusual behaviour in
2308 EVP_cmp_parameters() which returns 1 if the parameters match.
2309 This meant that parameters were omitted when they *didn't* match and
2310 the certificate was useless. Certificates signed with 'ca' didn't have
2312 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
2314 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
2315 The interface is as follows:
2316 Applications can use
2317 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
2318 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
2319 "off" is now the default.
2320 The library internally uses
2321 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
2322 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
2323 to disable memory-checking temporarily.
2325 Some inconsistent states that previously were possible (and were
2326 even the default) are now avoided.
2328 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
2329 with each memory chunk allocated; this is occasionally more helpful
2330 than just having a counter.
2332 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
2334 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
2338 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
2339 which largely parallels "options", but is for changing API behaviour,
2340 whereas "options" are about protocol behaviour.
2341 Initial "mode" flags are:
2343 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
2344 a single record has been written.
2345 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
2346 retries use the same buffer location.
2347 (But all of the contents must be
2351 *) Bugfix: SSL_set_mode ignored its parameter, only SSL_CTX_set_mode
2354 *) Fix problems with no-hmac etc.
2355 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
2357 *) New functions RSA_get_default_method(), RSA_set_method() and
2358 RSA_get_method(). These allows replacement of RSA_METHODs without having
2359 to mess around with the internals of an RSA structure.
2362 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
2363 Also really enable memory leak checks in openssl.c and in some
2365 [Chad C. Mulligan, Bodo Moeller]
2367 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
2368 up the length of negative integers. This has now been simplified to just
2369 store the length when it is first determined and use it later, rather
2370 than trying to keep track of where data is copied and updating it to
2372 [Steve Henson, reported by Brien Wheeler
2373 <bwheeler@authentica-security.com>]
2375 *) Add a new function PKCS7_signatureVerify. This allows the verification
2376 of a PKCS#7 signature but with the signing certificate passed to the
2377 function itself. This contrasts with PKCS7_dataVerify which assumes the
2378 certificate is present in the PKCS#7 structure. This isn't always the
2379 case: certificates can be omitted from a PKCS#7 structure and be
2380 distributed by "out of band" means (such as a certificate database).
2383 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
2384 function prototypes in pem.h, also change util/mkdef.pl to add the
2385 necessary function names.
2388 *) mk1mf.pl (used by Windows builds) did not properly read the
2389 options set by Configure in the top level Makefile, and Configure
2390 was not even able to write more than one option correctly.
2391 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
2394 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
2395 file to be loaded from a BIO or FILE pointer. The BIO version will
2396 for example allow memory BIOs to contain config info.
2399 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
2400 Whoever hopes to achieve shared-library compatibility across versions
2401 must use this, not the compile-time macro.
2402 (Exercise 0.9.4: Which is the minimum library version required by
2404 Note: All this applies only to multi-threaded programs, others don't
2408 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
2409 through a BIO pair triggered the default case, i.e.
2410 SSLerr(...,SSL_R_UNKNOWN_STATE).
2413 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
2414 can use the SSL library even if none of the specific BIOs is
2418 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
2419 for the encoded length.
2420 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
2422 *) Add initial documentation of the X509V3 functions.
2425 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
2426 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
2427 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
2428 secure PKCS#8 private key format with a high iteration count.
2431 *) Fix determination of Perl interpreter: A perl or perl5
2432 _directory_ in $PATH was also accepted as the interpreter.
2433 [Ralf S. Engelschall]
2435 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
2436 wrong with it but it was very old and did things like calling
2437 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
2441 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
2442 to use the new extension code.
2445 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
2446 with macros. This should make it easier to change their form, add extra
2447 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
2451 *) Add to configuration table a new entry that can specify an alternative
2452 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
2453 according to Mark Crispin <MRC@Panda.COM>.
2457 *) DES CBC did not update the IV. Weird.
2460 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
2461 Changing the behaviour of the former might break existing programs --
2462 where IV updating is needed, des_ncbc_encrypt can be used.
2465 *) When bntest is run from "make test" it drives bc to check its
2466 calculations, as well as internally checking them. If an internal check
2467 fails, it needs to cause bc to give a non-zero result or make test carries
2468 on without noticing the failure. Fixed.
2471 *) DES library cleanups.
2474 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
2475 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
2476 ciphers. NOTE: although the key derivation function has been verified
2477 against some published test vectors it has not been extensively tested
2478 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
2482 *) Instead of "mkdir -p", which is not fully portable, use new
2483 Perl script "util/mkdir-p.pl".
2486 *) Rewrite the way password based encryption (PBE) is handled. It used to
2487 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
2488 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
2489 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
2490 the 'parameter' field of the AlgorithmIdentifier is passed to the
2491 underlying key generation function so it must do its own ASN1 parsing.
2492 This has also changed the EVP_PBE_CipherInit() function which now has a
2493 'parameter' argument instead of literal salt and iteration count values
2494 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
2497 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
2498 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
2499 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
2500 KEY" because this clashed with PKCS#8 unencrypted string. Since this
2501 value was just used as a "magic string" and not used directly its
2502 value doesn't matter.
2505 *) Introduce some semblance of const correctness to BN. Shame C doesn't
2509 *) "linux-sparc64" configuration (ultrapenguin).
2510 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
2511 "linux-sparc" configuration.
2512 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
2514 *) config now generates no-xxx options for missing ciphers.
2517 *) Support the EBCDIC character set (work in progress).
2518 File ebcdic.c not yet included because it has a different license.
2519 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
2521 *) Support BS2000/OSD-POSIX.
2522 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
2524 *) Make callbacks for key generation use void * instead of char *.
2527 *) Make S/MIME samples compile (not yet tested).
2530 *) Additional typesafe stacks.
2533 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
2537 Changes between 0.9.3 and 0.9.3a [29 May 1999]
2539 *) New configuration variant "sco5-gcc".
2541 *) Updated some demos.
2542 [Sean O Riordain, Wade Scholine]
2544 *) Add missing BIO_free at exit of pkcs12 application.
2547 *) Fix memory leak in conf.c.
2550 *) Updates for Win32 to assembler version of MD5.
2553 *) Set #! path to perl in apps/der_chop to where we found it
2554 instead of using a fixed path.
2557 *) SHA library changes for irix64-mips4-cc.
2560 *) Improvements for VMS support.
2564 Changes between 0.9.2b and 0.9.3 [24 May 1999]
2566 *) Bignum library bug fix. IRIX 6 passes "make test" now!
2567 This also avoids the problems with SC4.2 and unpatched SC5.
2568 [Andy Polyakov <appro@fy.chalmers.se>]
2570 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
2571 These are required because of the typesafe stack would otherwise break
2572 existing code. If old code used a structure member which used to be STACK
2573 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
2574 sk_num or sk_value it would produce an error because the num, data members
2575 are not present in STACK_OF. Now it just produces a warning. sk_set
2576 replaces the old method of assigning a value to sk_value
2577 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
2578 that does this will no longer work (and should use sk_set instead) but
2579 this could be regarded as a "questionable" behaviour anyway.
2582 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
2583 correctly handle encrypted S/MIME data.
2586 *) Change type of various DES function arguments from des_cblock
2587 (which means, in function argument declarations, pointer to char)
2588 to des_cblock * (meaning pointer to array with 8 char elements),
2589 which allows the compiler to do more typechecking; it was like
2590 that back in SSLeay, but with lots of ugly casts.
2592 Introduce new type const_des_cblock.
2595 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
2596 problems: find RecipientInfo structure that matches recipient certificate
2597 and initialise the ASN1 structures properly based on passed cipher.
2600 *) Belatedly make the BN tests actually check the results.
2603 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
2604 to and from BNs: it was completely broken. New compilation option
2605 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
2606 key elements as negative integers.
2609 *) Reorganize and speed up MD5.
2610 [Andy Polyakov <appro@fy.chalmers.se>]
2613 [Richard Levitte <richard@levitte.org>]
2615 *) New option -out to asn1parse to allow the parsed structure to be
2616 output to a file. This is most useful when combined with the -strparse
2617 option to examine the output of things like OCTET STRINGS.
2620 *) Make SSL library a little more fool-proof by not requiring any longer
2621 that SSL_set_{accept,connect}_state be called before
2622 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
2623 in many applications because usually everything *appeared* to work as
2624 intended anyway -- now it really works as intended).
2627 *) Move openssl.cnf out of lib/.
2630 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
2631 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
2632 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
2633 [Ralf S. Engelschall]
2635 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
2636 handle PKCS#7 enveloped data properly.
2637 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
2639 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
2640 copying pointers. The cert_st handling is changed by this in
2641 various ways (and thus what used to be known as ctx->default_cert
2642 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
2643 any longer when s->cert does not give us what we need).
2644 ssl_cert_instantiate becomes obsolete by this change.
2645 As soon as we've got the new code right (possibly it already is?),
2646 we have solved a couple of bugs of the earlier code where s->cert
2647 was used as if it could not have been shared with other SSL structures.
2649 Note that using the SSL API in certain dirty ways now will result
2650 in different behaviour than observed with earlier library versions:
2651 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
2652 does not influence s as it used to.
2654 In order to clean up things more thoroughly, inside SSL_SESSION
2655 we don't use CERT any longer, but a new structure SESS_CERT
2656 that holds per-session data (if available); currently, this is
2657 the peer's certificate chain and, for clients, the server's certificate
2658 and temporary key. CERT holds only those values that can have
2659 meaningful defaults in an SSL_CTX.
2662 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
2663 from the internal representation. Various PKCS#7 fixes: remove some
2664 evil casts and set the enc_dig_alg field properly based on the signing
2668 *) Allow PKCS#12 password to be set from the command line or the
2669 environment. Let 'ca' get its config file name from the environment
2670 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
2674 *) Allow certificate policies extension to use an IA5STRING for the
2675 organization field. This is contrary to the PKIX definition but
2676 VeriSign uses it and IE5 only recognises this form. Document 'x509'
2680 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
2681 without disallowing inline assembler and the like for non-pedantic builds.
2684 *) Support Borland C++ builder.
2685 [Janez Jere <jj@void.si>, modified by Ulf Möller]
2690 *) SHA-1 cleanups and performance enhancements.
2691 [Andy Polyakov <appro@fy.chalmers.se>]
2693 *) Sparc v8plus assembler for the bignum library.
2694 [Andy Polyakov <appro@fy.chalmers.se>]
2696 *) Accept any -xxx and +xxx compiler options in Configure.
2699 *) Update HPUX configuration.
2702 *) Add missing sk_<type>_unshift() function to safestack.h
2703 [Ralf S. Engelschall]
2705 *) New function SSL_CTX_use_certificate_chain_file that sets the
2706 "extra_cert"s in addition to the certificate. (This makes sense
2707 only for "PEM" format files, as chains as a whole are not
2711 *) Support verify_depth from the SSL API.
2712 x509_vfy.c had what can be considered an off-by-one-error:
2713 Its depth (which was not part of the external interface)
2714 was actually counting the number of certificates in a chain;
2715 now it really counts the depth.
2718 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
2719 instead of X509err, which often resulted in confusing error
2720 messages since the error codes are not globally unique
2721 (e.g. an alleged error in ssl3_accept when a certificate
2722 didn't match the private key).
2724 *) New function SSL_CTX_set_session_id_context that allows to set a default
2725 value (so that you don't need SSL_set_session_id_context for each
2726 connection using the SSL_CTX).
2729 *) OAEP decoding bug fix.
2732 *) Support INSTALL_PREFIX for package builders, as proposed by
2736 *) New Configure options "threads" and "no-threads". For systems
2737 where the proper compiler options are known (currently Solaris
2738 and Linux), "threads" is the default.
2741 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
2744 *) Install various scripts to $(OPENSSLDIR)/misc, not to
2745 $(INSTALLTOP)/bin -- they shouldn't clutter directories
2746 such as /usr/local/bin.
2749 *) "make linux-shared" to build shared libraries.
2750 [Niels Poppe <niels@netbox.org>]
2752 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
2755 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
2756 extension adding in x509 utility.
2759 *) Remove NOPROTO sections and error code comments.
2762 *) Partial rewrite of the DEF file generator to now parse the ANSI
2766 *) New Configure options --prefix=DIR and --openssldir=DIR.
2769 *) Complete rewrite of the error code script(s). It is all now handled
2770 by one script at the top level which handles error code gathering,
2771 header rewriting and C source file generation. It should be much better
2772 than the old method: it now uses a modified version of Ulf's parser to
2773 read the ANSI prototypes in all header files (thus the old K&R definitions
2774 aren't needed for error creation any more) and do a better job of
2775 translating function codes into names. The old 'ASN1 error code imbedded
2776 in a comment' is no longer necessary and it doesn't use .err files which
2777 have now been deleted. Also the error code call doesn't have to appear all
2778 on one line (which resulted in some large lines...).
2781 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
2784 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
2785 0 (which usually indicates a closed connection), but continue reading.
2788 *) Fix some race conditions.
2791 *) Add support for CRL distribution points extension. Add Certificate
2792 Policies and CRL distribution points documentation.
2795 *) Move the autogenerated header file parts to crypto/opensslconf.h.
2798 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
2799 8 of keying material. Merlin has also confirmed interop with this fix
2800 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
2801 [Merlin Hughes <merlin@baltimore.ie>]
2803 *) Fix lots of warnings.
2804 [Richard Levitte <levitte@stacken.kth.se>]
2806 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
2807 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
2808 [Richard Levitte <levitte@stacken.kth.se>]
2810 *) Fix problems with sizeof(long) == 8.
2811 [Andy Polyakov <appro@fy.chalmers.se>]
2813 *) Change functions to ANSI C.
2816 *) Fix typos in error codes.
2817 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
2819 *) Remove defunct assembler files from Configure.
2822 *) SPARC v8 assembler BIGNUM implementation.
2823 [Andy Polyakov <appro@fy.chalmers.se>]
2825 *) Support for Certificate Policies extension: both print and set.
2826 Various additions to support the r2i method this uses.
2829 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
2830 return a const string when you are expecting an allocated buffer.
2833 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
2834 types DirectoryString and DisplayText.
2837 *) Add code to allow r2i extensions to access the configuration database,
2838 add an LHASH database driver and add several ctx helper functions.
2841 *) Fix an evil bug in bn_expand2() which caused various BN functions to
2842 fail when they extended the size of a BIGNUM.
2845 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
2846 support typesafe stack.
2849 *) Fix typo in SSL_[gs]et_options().
2850 [Nils Frostberg <nils@medcom.se>]
2852 *) Delete various functions and files that belonged to the (now obsolete)
2853 old X509V3 handling code.
2856 *) New Configure option "rsaref".
2859 *) Don't auto-generate pem.h.
2862 *) Introduce type-safe ASN.1 SETs.
2865 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
2866 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
2868 *) Introduce type-safe STACKs. This will almost certainly break lots of code
2869 that links with OpenSSL (well at least cause lots of warnings), but fear
2870 not: the conversion is trivial, and it eliminates loads of evil casts. A
2871 few STACKed things have been converted already. Feel free to convert more.
2872 In the fullness of time, I'll do away with the STACK type altogether.
2875 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
2876 specified in <certfile> by updating the entry in the index.txt file.
2877 This way one no longer has to edit the index.txt file manually for
2878 revoking a certificate. The -revoke option does the gory details now.
2879 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
2881 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
2882 `-text' option at all and this way the `-noout -text' combination was
2883 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
2884 [Ralf S. Engelschall]
2886 *) Make sure a corresponding plain text error message exists for the
2887 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
2888 verify callback function determined that a certificate was revoked.
2889 [Ralf S. Engelschall]
2891 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
2892 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
2893 all available cipers including rc5, which was forgotten until now.
2894 In order to let the testing shell script know which algorithms
2895 are available, a new (up to now undocumented) command
2896 "openssl list-cipher-commands" is used.
2899 *) Bugfix: s_client occasionally would sleep in select() when
2900 it should have checked SSL_pending() first.
2903 *) New functions DSA_do_sign and DSA_do_verify to provide access to
2904 the raw DSA values prior to ASN.1 encoding.
2907 *) Tweaks to Configure
2908 [Niels Poppe <niels@netbox.org>]
2910 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
2914 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
2917 *) New config option to avoid instructions that are illegal on the 80386.
2918 The default code is faster, but requires at least a 486.
2921 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
2922 SSL2_SERVER_VERSION (not used at all) macros, which are now the
2923 same as SSL2_VERSION anyway.
2926 *) New "-showcerts" option for s_client.
2929 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
2930 application. Various cleanups and fixes.
2933 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
2934 modify error routines to work internally. Add error codes and PBE init
2935 to library startup routines.
2938 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
2939 packing functions to asn1 and evp. Changed function names and error
2940 codes along the way.
2943 *) PKCS12 integration: and so it begins... First of several patches to
2944 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
2945 objects to objects.h
2948 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
2949 and display support for Thawte strong extranet extension.
2952 *) Add LinuxPPC support.
2953 [Jeff Dubrule <igor@pobox.org>]
2955 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
2956 bn_div_words in alpha.s.
2957 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
2959 *) Make sure the RSA OAEP test is skipped under -DRSAref because
2960 OAEP isn't supported when OpenSSL is built with RSAref.
2961 [Ulf Moeller <ulf@fitug.de>]
2963 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
2964 so they no longer are missing under -DNOPROTO.
2965 [Soren S. Jorvang <soren@t.dk>]
2968 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
2970 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
2971 doesn't work when the session is reused. Coming soon!
2974 *) Fix a security hole, that allows sessions to be reused in the wrong
2975 context thus bypassing client cert protection! All software that uses
2976 client certs and session caches in multiple contexts NEEDS PATCHING to
2977 allow session reuse! A fuller solution is in the works.
2978 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
2980 *) Some more source tree cleanups (removed obsolete files
2981 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
2982 permission on "config" script to be executable) and a fix for the INSTALL
2984 [Ulf Moeller <ulf@fitug.de>]
2986 *) Remove some legacy and erroneous uses of malloc, free instead of
2988 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
2990 *) Make rsa_oaep_test return non-zero on error.
2991 [Ulf Moeller <ulf@fitug.de>]
2993 *) Add support for native Solaris shared libraries. Configure
2994 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
2995 if someone would make that last step automatic.
2996 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
2998 *) ctx_size was not built with the right compiler during "make links". Fixed.
3001 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
3002 except NULL ciphers". This means the default cipher list will no longer
3003 enable NULL ciphers. They need to be specifically enabled e.g. with
3004 the string "DEFAULT:eNULL".
3007 *) Fix to RSA private encryption routines: if p < q then it would
3008 occasionally produce an invalid result. This will only happen with
3009 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
3012 *) Be less restrictive and allow also `perl util/perlpath.pl
3013 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
3014 because this way one can also use an interpreter named `perl5' (which is
3015 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
3016 installed as `perl').
3017 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
3019 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
3020 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
3022 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
3023 advapi32.lib to Win32 build and change the pem test comparision
3024 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
3025 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
3026 and crypto/des/ede_cbcm_enc.c.
3029 *) DES quad checksum was broken on big-endian architectures. Fixed.
3032 *) Comment out two functions in bio.h that aren't implemented. Fix up the
3033 Win32 test batch file so it (might) work again. The Win32 test batch file
3034 is horrible: I feel ill....
3037 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
3038 in e_os.h. Audit of header files to check ANSI and non ANSI
3039 sections: 10 functions were absent from non ANSI section and not exported
3040 from Windows DLLs. Fixed up libeay.num for new functions.
3043 *) Make `openssl version' output lines consistent.
3044 [Ralf S. Engelschall]
3046 *) Fix Win32 symbol export lists for BIO functions: Added
3047 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
3048 to ms/libeay{16,32}.def.
3049 [Ralf S. Engelschall]
3051 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
3052 fine under Unix and passes some trivial tests I've now added. But the
3053 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
3054 added to make sure no one expects that this stuff really works in the
3055 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
3056 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
3058 [Ralf S. Engelschall]
3060 *) Fix the generation of two part addresses in perl.
3061 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
3063 *) Add config entry for Linux on MIPS.
3064 [John Tobey <jtobey@channel1.com>]
3066 *) Make links whenever Configure is run, unless we are on Windoze.
3069 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
3070 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
3074 *) Add a useful kludge to allow package maintainers to specify compiler and
3075 other platforms details on the command line without having to patch the
3076 Configure script everytime: One now can use ``perl Configure
3077 <id>:<details>'', i.e. platform ids are allowed to have details appended
3078 to them (seperated by colons). This is treated as there would be a static
3079 pre-configured entry in Configure's %table under key <id> with value
3080 <details> and ``perl Configure <id>'' is called. So, when you want to
3081 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
3082 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
3083 now, which overrides the FreeBSD-elf entry on-the-fly.
3084 [Ralf S. Engelschall]
3086 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
3089 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
3090 on the `perl Configure ...' command line. This way one can compile
3091 OpenSSL libraries with Position Independent Code (PIC) which is needed
3092 for linking it into DSOs.
3093 [Ralf S. Engelschall]
3095 *) Remarkably, export ciphers were totally broken and no-one had noticed!
3099 *) Cleaned up the LICENSE document: The official contact for any license
3100 questions now is the OpenSSL core team under openssl-core@openssl.org.
3101 And add a paragraph about the dual-license situation to make sure people
3102 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
3103 to the OpenSSL toolkit.
3104 [Ralf S. Engelschall]
3106 *) General source tree makefile cleanups: Made `making xxx in yyy...'
3107 display consistent in the source tree and replaced `/bin/rm' by `rm'.
3108 Additonally cleaned up the `make links' target: Remove unnecessary
3109 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
3110 to speed processing and no longer clutter the display with confusing
3111 stuff. Instead only the actually done links are displayed.
3112 [Ralf S. Engelschall]
3114 *) Permit null encryption ciphersuites, used for authentication only. It used
3115 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
3116 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
3120 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
3121 signed attributes when verifying signatures (this would break them),
3122 the detached data encoding was wrong and public keys obtained using
3123 X509_get_pubkey() weren't freed.
3126 *) Add text documentation for the BUFFER functions. Also added a work around
3127 to a Win95 console bug. This was triggered by the password read stuff: the
3128 last character typed gets carried over to the next fread(). If you were
3129 generating a new cert request using 'req' for example then the last
3130 character of the passphrase would be CR which would then enter the first
3134 *) Added the new `Includes OpenSSL Cryptography Software' button as
3135 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
3136 button and can be used by applications based on OpenSSL to show the
3137 relationship to the OpenSSL project.
3138 [Ralf S. Engelschall]
3140 *) Remove confusing variables in function signatures in files
3141 ssl/ssl_lib.c and ssl/ssl.h.
3142 [Lennart Bong <lob@kulthea.stacken.kth.se>]
3144 *) Don't install bss_file.c under PREFIX/include/
3145 [Lennart Bong <lob@kulthea.stacken.kth.se>]
3147 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
3148 functions that return function pointers and has support for NT specific
3149 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
3150 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
3151 unsigned to signed types: this was killing the Win32 compile.
3154 *) Add new certificate file to stack functions,
3155 SSL_add_dir_cert_subjects_to_stack() and
3156 SSL_add_file_cert_subjects_to_stack(). These largely supplant
3157 SSL_load_client_CA_file(), and can be used to add multiple certs easily
3158 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
3159 This means that Apache-SSL and similar packages don't have to mess around
3160 to add as many CAs as they want to the preferred list.
3163 *) Experiment with doxygen documentation. Currently only partially applied to
3165 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
3166 openssl.doxy as the configuration file.
3169 *) Get rid of remaining C++-style comments which strict C compilers hate.
3170 [Ralf S. Engelschall, pointed out by Carlos Amengual]
3172 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
3173 compiled in by default: it has problems with large keys.
3176 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
3177 DH private keys and/or callback functions which directly correspond to
3178 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
3179 is needed for applications which have to configure certificates on a
3180 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
3182 For the RSA certificate situation is makes no difference, but
3183 for the DSA certificate situation this fixes the "no shared cipher"
3184 problem where the OpenSSL cipher selection procedure failed because the
3185 temporary keys were not overtaken from the context and the API provided
3186 no way to reconfigure them.
3187 The new functions now let applications reconfigure the stuff and they
3188 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
3189 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
3190 non-public-API function ssl_cert_instantiate() is used as a helper
3191 function and also to reduce code redundancy inside ssl_rsa.c.
3192 [Ralf S. Engelschall]
3194 *) Move s_server -dcert and -dkey options out of the undocumented feature
3195 area because they are useful for the DSA situation and should be
3196 recognized by the users.
3197 [Ralf S. Engelschall]
3199 *) Fix the cipher decision scheme for export ciphers: the export bits are
3200 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
3201 SSL_EXP_MASK. So, the original variable has to be used instead of the
3202 already masked variable.
3203 [Richard Levitte <levitte@stacken.kth.se>]
3205 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
3206 [Richard Levitte <levitte@stacken.kth.se>]
3208 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
3209 from `int' to `unsigned int' because it's a length and initialized by
3210 EVP_DigestFinal() which expects an `unsigned int *'.
3211 [Richard Levitte <levitte@stacken.kth.se>]
3213 *) Don't hard-code path to Perl interpreter on shebang line of Configure
3214 script. Instead use the usual Shell->Perl transition trick.
3215 [Ralf S. Engelschall]
3217 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
3218 (in addition to RSA certificates) to match the behaviour of `openssl dsa
3219 -noout -modulus' as it's already the case for `openssl rsa -noout
3220 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
3221 currently the public key is printed (a decision which was already done by
3222 `openssl dsa -modulus' in the past) which serves a similar purpose.
3223 Additionally the NO_RSA no longer completely removes the whole -modulus
3224 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
3226 [Ralf S. Engelschall]
3228 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
3229 BIO. See the source (crypto/evp/bio_ok.c) for more info.
3230 [Arne Ansper <arne@ats.cyber.ee>]
3232 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
3233 to be added. Now both 'req' and 'ca' can use new objects defined in the
3237 *) Add cool BIO that does syslog (or event log on NT).
3238 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
3240 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
3241 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
3242 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
3243 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
3246 *) Add preliminary config info for new extension code.
3249 *) Make RSA_NO_PADDING really use no padding.
3250 [Ulf Moeller <ulf@fitug.de>]
3252 *) Generate errors when private/public key check is done.
3255 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
3256 for some CRL extensions and new objects added.
3259 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
3260 key usage extension and fuller support for authority key id.
3263 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
3264 padding method for RSA, which is recommended for new applications in PKCS
3265 #1 v2.0 (RFC 2437, October 1998).
3266 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
3267 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
3268 against Bleichbacher's attack on RSA.
3269 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
3272 *) Updates to the new SSL compression code
3273 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
3275 *) Fix so that the version number in the master secret, when passed
3276 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
3277 (because the server will not accept higher), that the version number
3278 is 0x03,0x01, not 0x03,0x00
3279 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
3281 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
3282 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
3283 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
3286 *) Support for RAW extensions where an arbitrary extension can be
3287 created by including its DER encoding. See apps/openssl.cnf for
3291 *) Make sure latest Perl versions don't interpret some generated C array
3292 code as Perl array code in the crypto/err/err_genc.pl script.
3293 [Lars Weber <3weber@informatik.uni-hamburg.de>]
3295 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
3296 not many people have the assembler. Various Win32 compilation fixes and
3297 update to the INSTALL.W32 file with (hopefully) more accurate Win32
3301 *) Modify configure script 'Configure' to automatically create crypto/date.h
3302 file under Win32 and also build pem.h from pem.org. New script
3303 util/mkfiles.pl to create the MINFO file on environments that can't do a
3304 'make files': perl util/mkfiles.pl >MINFO should work.
3307 *) Major rework of DES function declarations, in the pursuit of correctness
3308 and purity. As a result, many evil casts evaporated, and some weirdness,
3309 too. You may find this causes warnings in your code. Zapping your evil
3310 casts will probably fix them. Mostly.
3313 *) Fix for a typo in asn1.h. Bug fix to object creation script
3314 obj_dat.pl. It considered a zero in an object definition to mean
3315 "end of object": none of the objects in objects.h have any zeros
3316 so it wasn't spotted.
3317 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
3319 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
3320 Masking (CBCM). In the absence of test vectors, the best I have been able
3321 to do is check that the decrypt undoes the encrypt, so far. Send me test
3322 vectors if you have them.
3325 *) Correct calculation of key length for export ciphers (too much space was
3326 allocated for null ciphers). This has not been tested!
3329 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
3330 message is now correct (it understands "crypto" and "ssl" on its
3331 command line). There is also now an "update" option. This will update
3332 the util/ssleay.num and util/libeay.num files with any new functions.
3334 perl util/mkdef.pl crypto ssl update
3335 it will update them.
3338 *) Overhauled the Perl interface (perl/*):
3339 - ported BN stuff to OpenSSL's different BN library
3340 - made the perl/ source tree CVS-aware
3341 - renamed the package from SSLeay to OpenSSL (the files still contain
3342 their history because I've copied them in the repository)
3343 - removed obsolete files (the test scripts will be replaced
3344 by better Test::Harness variants in the future)
3345 [Ralf S. Engelschall]
3347 *) First cut for a very conservative source tree cleanup:
3348 1. merge various obsolete readme texts into doc/ssleay.txt
3349 where we collect the old documents and readme texts.
3350 2. remove the first part of files where I'm already sure that we no
3351 longer need them because of three reasons: either they are just temporary
3352 files which were left by Eric or they are preserved original files where
3353 I've verified that the diff is also available in the CVS via "cvs diff
3354 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
3355 the crypto/md/ stuff).
3356 [Ralf S. Engelschall]
3358 *) More extension code. Incomplete support for subject and issuer alt
3359 name, issuer and authority key id. Change the i2v function parameters
3360 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
3361 what that's for :-) Fix to ASN1 macro which messed up
3362 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
3365 *) Preliminary support for ENUMERATED type. This is largely copied from the
3369 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
3370 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
3372 *) Make sure `make rehash' target really finds the `openssl' program.
3373 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
3375 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
3376 like to hear about it if this slows down other processors.
3379 *) Add CygWin32 platform information to Configure script.
3380 [Alan Batie <batie@aahz.jf.intel.com>]
3382 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
3383 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
3385 *) New program nseq to manipulate netscape certificate sequences
3388 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
3392 *) Fixes to BN code. Previously the default was to define BN_RECURSION
3393 but the BN code had some problems that would cause failures when
3394 doing certificate verification and some other functions.
3395 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
3397 *) Add ASN1 and PEM code to support netscape certificate sequences.
3400 *) Add ASN1 and PEM code to support netscape certificate sequences.
3403 *) Add several PKIX and private extended key usage OIDs.
3406 *) Modify the 'ca' program to handle the new extension code. Modify
3407 openssl.cnf for new extension format, add comments.
3410 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
3411 and add a sample to openssl.cnf so req -x509 now adds appropriate
3415 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
3416 error code, add initial support to X509_print() and x509 application.
3419 *) Takes a deep breath and start addding X509 V3 extension support code. Add
3420 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
3421 stuff is currently isolated and isn't even compiled yet.
3424 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
3425 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
3426 Removed the versions check from X509 routines when loading extensions:
3427 this allows certain broken certificates that don't set the version
3428 properly to be processed.
3431 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
3432 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
3433 can still be regenerated with "make depend".
3436 *) Spelling mistake in C version of CAST-128.
3437 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
3439 *) Changes to the error generation code. The perl script err-code.pl
3440 now reads in the old error codes and retains the old numbers, only
3441 adding new ones if necessary. It also only changes the .err files if new
3442 codes are added. The makefiles have been modified to only insert errors
3443 when needed (to avoid needlessly modifying header files). This is done
3444 by only inserting errors if the .err file is newer than the auto generated
3445 C file. To rebuild all the error codes from scratch (the old behaviour)
3446 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
3447 or delete all the .err files.
3450 *) CAST-128 was incorrectly implemented for short keys. The C version has
3451 been fixed, but is untested. The assembler versions are also fixed, but
3452 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
3453 to regenerate it if needed.
3454 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
3455 Hagino <itojun@kame.net>]
3457 *) File was opened incorrectly in randfile.c.
3458 [Ulf Möller <ulf@fitug.de>]
3460 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
3461 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
3462 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
3463 al: it's just almost always a UTCTime. Note this patch adds new error
3464 codes so do a "make errors" if there are problems.
3467 *) Correct Linux 1 recognition in config.
3468 [Ulf Möller <ulf@fitug.de>]
3470 *) Remove pointless MD5 hash when using DSA keys in ca.
3471 [Anonymous <nobody@replay.com>]
3473 *) Generate an error if given an empty string as a cert directory. Also
3474 generate an error if handed NULL (previously returned 0 to indicate an
3475 error, but didn't set one).
3476 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
3478 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
3481 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
3482 parameters. This was causing a warning which killed off the Win32 compile.
3485 *) Remove C++ style comments from crypto/bn/bn_local.h.
3486 [Neil Costigan <neil.costigan@celocom.com>]
3488 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
3489 based on a text string, looking up short and long names and finally
3490 "dot" format. The "dot" format stuff didn't work. Added new function
3491 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
3492 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
3493 OID is not part of the table.
3496 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
3497 X509_LOOKUP_by_alias().
3500 *) Sort openssl functions by name.
3503 *) Get the gendsa program working (hopefully) and add it to app list. Remove
3504 encryption from sample DSA keys (in case anyone is interested the password
3508 *) Make _all_ *_free functions accept a NULL pointer.
3509 [Frans Heymans <fheymans@isaserver.be>]
3511 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
3513 [Anonymous <nobody@replay.com>]
3515 *) s_server should send the CAfile as acceptable CAs, not its own cert.
3516 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
3518 *) Don't blow it for numeric -newkey arguments to apps/req.
3519 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
3521 *) Temp key "for export" tests were wrong in s3_srvr.c.
3522 [Anonymous <nobody@replay.com>]
3524 *) Add prototype for temp key callback functions
3525 SSL_CTX_set_tmp_{rsa,dh}_callback().
3528 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
3529 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
3532 *) X509_name_add_entry() freed the wrong thing after an error.
3533 [Arne Ansper <arne@ats.cyber.ee>]
3535 *) rsa_eay.c would attempt to free a NULL context.
3536 [Arne Ansper <arne@ats.cyber.ee>]
3538 *) BIO_s_socket() had a broken should_retry() on Windoze.
3539 [Arne Ansper <arne@ats.cyber.ee>]
3541 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
3542 [Arne Ansper <arne@ats.cyber.ee>]
3544 *) Make sure the already existing X509_STORE->depth variable is initialized
3545 in X509_STORE_new(), but document the fact that this variable is still
3546 unused in the certificate verification process.
3547 [Ralf S. Engelschall]
3549 *) Fix the various library and apps files to free up pkeys obtained from
3550 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
3553 *) Fix reference counting in X509_PUBKEY_get(). This makes
3554 demos/maurice/example2.c work, amongst others, probably.
3555 [Steve Henson and Ben Laurie]
3557 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
3558 `openssl' and second, the shortcut symlinks for the `openssl <command>'
3559 are no longer created. This way we have a single and consistent command
3560 line interface `openssl <command>', similar to `cvs <command>'.
3561 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
3563 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
3564 BIT STRING wrapper always have zero unused bits.
3567 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
3570 *) Make the top-level INSTALL documentation easier to understand.
3573 *) Makefiles updated to exit if an error occurs in a sub-directory
3574 make (including if user presses ^C) [Paul Sutton]
3576 *) Make Montgomery context stuff explicit in RSA data structure.
3579 *) Fix build order of pem and err to allow for generated pem.h.
3582 *) Fix renumbering bug in X509_NAME_delete_entry().
3585 *) Enhanced the err-ins.pl script so it makes the error library number
3586 global and can add a library name. This is needed for external ASN1 and
3587 other error libraries.
3590 *) Fixed sk_insert which never worked properly.
3593 *) Fix ASN1 macros so they can handle indefinite length construted
3594 EXPLICIT tags. Some non standard certificates use these: they can now
3598 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
3599 into a single doc/ssleay.txt bundle. This way the information is still
3600 preserved but no longer messes up this directory. Now it's new room for
3601 the new set of documenation files.
3602 [Ralf S. Engelschall]
3604 *) SETs were incorrectly DER encoded. This was a major pain, because they
3605 shared code with SEQUENCEs, which aren't coded the same. This means that
3606 almost everything to do with SETs or SEQUENCEs has either changed name or
3607 number of arguments.
3608 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
3610 *) Fix test data to work with the above.
3613 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
3614 was already fixed by Eric for 0.9.1 it seems.
3615 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
3617 *) Autodetect FreeBSD3.
3620 *) Fix various bugs in Configure. This affects the following platforms:
3624 unixware-2.0-pentium
3628 *) Eliminate generated files from CVS. Reorder tests to regenerate files
3629 before they are needed.
3632 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
3636 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
3638 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
3639 changed SSLeay to OpenSSL in version strings.
3640 [Ralf S. Engelschall]
3642 *) Some fixups to the top-level documents.
3645 *) Fixed the nasty bug where rsaref.h was not found under compile-time
3646 because the symlink to include/ was missing.
3647 [Ralf S. Engelschall]
3649 *) Incorporated the popular no-RSA/DSA-only patches
3650 which allow to compile a RSA-free SSLeay.
3651 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
3653 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
3654 when "ssleay" is still not found.
3655 [Ralf S. Engelschall]
3657 *) Added more platforms to Configure: Cray T3E, HPUX 11,
3658 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
3660 *) Updated the README file.
3661 [Ralf S. Engelschall]
3663 *) Added various .cvsignore files in the CVS repository subdirs
3664 to make a "cvs update" really silent.
3665 [Ralf S. Engelschall]
3667 *) Recompiled the error-definition header files and added
3668 missing symbols to the Win32 linker tables.
3669 [Ralf S. Engelschall]
3671 *) Cleaned up the top-level documents;
3672 o new files: CHANGES and LICENSE
3673 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
3674 o merged COPYRIGHT into LICENSE
3675 o removed obsolete TODO file
3676 o renamed MICROSOFT to INSTALL.W32
3677 [Ralf S. Engelschall]
3679 *) Removed dummy files from the 0.9.1b source tree:
3680 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
3681 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
3682 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
3683 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
3684 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
3685 [Ralf S. Engelschall]
3687 *) Added various platform portability fixes.
3690 *) The Genesis of the OpenSSL rpject:
3691 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
3692 Young and Tim J. Hudson created while they were working for C2Net until
3694 [The OpenSSL Project]
3697 Changes between 0.9.0b and 0.9.1b [not released]
3699 *) Updated a few CA certificates under certs/
3702 *) Changed some BIGNUM api stuff.
3705 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
3706 DGUX x86, Linux Alpha, etc.
3709 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
3710 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
3714 *) Add -strparse option to asn1pars program which parses nested
3716 [Dr Stephen Henson <shenson@bigfoot.com>]
3718 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
3721 *) DSA fix for "ca" program.
3724 *) Added "-genkey" option to "dsaparam" program.
3727 *) Added RIPE MD160 (rmd160) message digest.
3730 *) Added -a (all) option to "ssleay version" command.
3733 *) Added PLATFORM define which is the id given to Configure.
3736 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
3739 *) Extended the ASN.1 parser routines.
3742 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
3745 *) Added a BN_CTX to the BN library.
3748 *) Fixed the weak key values in DES library
3751 *) Changed API in EVP library for cipher aliases.
3754 *) Added support for RC2/64bit cipher.
3757 *) Converted the lhash library to the crypto/mem.c functions.
3760 *) Added more recognized ASN.1 object ids.
3763 *) Added more RSA padding checks for SSL/TLS.
3766 *) Added BIO proxy/filter functionality.
3769 *) Added extra_certs to SSL_CTX which can be used
3770 send extra CA certificates to the client in the CA cert chain sending
3771 process. It can be configured with SSL_CTX_add_extra_chain_cert().
3774 *) Now Fortezza is denied in the authentication phase because
3775 this is key exchange mechanism is not supported by SSLeay at all.
3778 *) Additional PKCS1 checks.
3781 *) Support the string "TLSv1" for all TLS v1 ciphers.
3784 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
3785 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
3788 *) Fixed a few memory leaks.
3791 *) Fixed various code and comment typos.
3794 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
3795 bytes sent in the client random.
3796 [Edward Bishop <ebishop@spyglass.com>]
projects / openssl.git / blob