openssl.git
23 years agoVMS was behind when it comes to OCSP.
Richard Levitte [Fri, 20 Apr 2001 12:37:14 +0000 (12:37 +0000)]
VMS was behind when it comes to OCSP.

23 years agoAdd notes about the new ENGINE functionality.
Geoff Thorpe [Thu, 19 Apr 2001 23:06:00 +0000 (23:06 +0000)]
Add notes about the new ENGINE functionality.

23 years agoChanges to "openssl engine" to support the new control command code in
Geoff Thorpe [Thu, 19 Apr 2001 02:08:26 +0000 (02:08 +0000)]
Changes to "openssl engine" to support the new control command code in
ENGINE.

 * Extra verbosity can be added with more "v"'s, eg. '-vvv' gives
   information about input flags and descriptions for each control command
   in each ENGINE. Check the output of "openssl engine -vvv" for example.

 * '-pre <cmd>' and '-post <cmd>' can be used to invoke control commands on
   the specified ENGINE (or on all of them if no engine id is specified,
   although that usually gets pretty ugly). '-post' commands are only
   attempted if '-t' is specified and the engine successfully initialises.
   '-pre' commands are always attempted whether or not '-t' causes an
   initialisation to be tried afterwards. Multiple '-pre' and/or '-post'
   commands can be specified and they will be called in the order they
   occur on the command line.

Parameterised commands (the normal case, there are currently no
unparameterised ones) are split into command and argument via a separating
colon. Eg. "openssl engine -pre SO_PATH:/lib/libdriver.so <id>" results in
the call;
    ENGINE_ctrl_cmd_string(e, "SO_PATH", "/lib/libdriver.so", 0);

Application code should similarly allow arbitrary name-value string pairs
to be passed into ENGINEs in a manner matching that in apps/engine.c,
either using the same colon-separated format, or entered as two distinct
strings. Eg. as stored in a registry. The last parameter of
ENGINE_ctrl_cmd_string can be changed from 0 to 1 if the command should
only be attempted if it's supported by the specified ENGINE (eg. for
commands like "FORK_CHECK:1" that may or may not apply to the run-time
ENGINE).

23 years agoSome more tweaks to ENGINE code.
Geoff Thorpe [Thu, 19 Apr 2001 01:45:40 +0000 (01:45 +0000)]
Some more tweaks to ENGINE code.

This change adds some basic control commands to the existing ENGINEs
(except the software 'openssl' engine). All these engines currently load
shared-libraries for hardware APIs, so they've all been given "SO_PATH"
commands that will configure the chosen ENGINE to load its shared library
from the given path. Eg. by calling;
    ENGINE_ctrl_cmd_string(e, "SO_PATH", <path>, 0).

The nCipher 'chil' ENGINE has also had "FORK_CHECK" and "THREAD_LOCKING"
commands added so these settings could be handled via application-level
configuration rather than in application source code.

Changes to "openssl engine" to test and examine these control commands will
be made shortly. It will also provide the necessary tips to application
programs wanting to support these dynamic control commands.

23 years agoSome BIG tweaks to ENGINE code.
Geoff Thorpe [Thu, 19 Apr 2001 00:41:55 +0000 (00:41 +0000)]
Some BIG tweaks to ENGINE code.

This change adds some new functionality to the ENGINE code and API to
make it possible for ENGINEs to describe and implement their own control
commands that can be interrogated and used by calling applications at
run-time. The source code includes numerous comments explaining how it all
works and some of the finer details. But basically, an ENGINE will normally
declare an array of ENGINE_CMD_DEFN entries in its ENGINE - and the various
new ENGINE_CTRL_*** command types take care of iterating through this list
of definitions, converting command numbers to names, command names to
numbers, getting descriptions, getting input flags, etc. These
administrative commands are handled directly in the base ENGINE code rather
than in each ENGINE's ctrl() handler, unless they specify the
ENGINE_FLAGS_MANUAL_CMD_CTRL flag (ie. if they're doing something clever or
dynamic with the command definitions).

There is also a new function, ENGINE_cmd_is_executable(), that will
determine if an ENGINE control command is of an "executable" type that
can be used in another new function, ENGINE_ctrl_cmd_string(). If not, the
control command is not supposed to be exposed out to user/config level
access - eg. it could involve the exchange of binary data, returning
results to calling code, etc etc. If the command is executable then
ENGINE_ctrl_cmd_string() can be called using a name/arg string pair. The
control command's input flags will be used to determine necessary
conversions before the control command is called, and commands of this
form will always return zero or one (failure or success, respectively).
This is set up so that arbitrary applications can support control commands
in a consistent way so that tweaking particular ENGINE behaviour is
specific to the ENGINE and the host environment, and independant of the
application or OpenSSL.

Some code demonstrating this stuff in action will applied shortly to the
various ENGINE implementations, as well as "openssl engine" support for
executing arbitrary control commands before and/or after initialising
various ENGINEs.

23 years agoSome more tweaks to ENGINE code.
Geoff Thorpe [Wed, 18 Apr 2001 21:46:00 +0000 (21:46 +0000)]
Some more tweaks to ENGINE code.

The existing ENGINEs (including the default 'openssl' software engine) were
static, declared inside the source file for each engine implementation. The
reason this was not going boom was that all the ENGINEs had reference
counts that never hit zero (once linked into the internal list, each would
always have at least 1 lasting structural reference).

To fix this so it will stay standing when an "unload" function is added to
match ENGINE_load_builtin_engines(), the "constructor" functions for each
ENGINE implementation have been changed to dynamically allocate and
construct their own ENGINEs using API functions. The other benefit of this
is that no ENGINE implementation has to include the internal "engine_int.h"
header file any more.

23 years agoMake a note of the recent ENGINE developments.
Geoff Thorpe [Wed, 18 Apr 2001 21:12:02 +0000 (21:12 +0000)]
Make a note of the recent ENGINE developments.

23 years agotypo
Bodo Möller [Wed, 18 Apr 2001 15:11:42 +0000 (15:11 +0000)]
typo

23 years agofix md_rand.c locking bugs
Bodo Möller [Wed, 18 Apr 2001 15:07:35 +0000 (15:07 +0000)]
fix md_rand.c locking bugs

23 years agoMake the shared library name and function symbol for the "nuron" ENGINE
Geoff Thorpe [Wed, 18 Apr 2001 04:47:01 +0000 (04:47 +0000)]
Make the shared library name and function symbol for the "nuron" ENGINE
static data where they could be parameterised by ctrl() commands.

23 years agoSome more tweaks from ENGINE code.
Geoff Thorpe [Wed, 18 Apr 2001 04:18:16 +0000 (04:18 +0000)]
Some more tweaks from ENGINE code.

Previously RAND_get_rand_method was returning a non-const pointer, but it
should be const. As with all other such cases, METHOD pointers are stored and
returned as "const". The only methods one should be able to alter are methods
"local" to the relevant code, in which case a non-const handle to the methods
should already exist.

This change has been forced by the constifying of the ENGINE code (before
which RAND_METHOD was the only method pointer in an ENGINE structure that was
not constant).

23 years agoSome more tweaks to ENGINE code.
Geoff Thorpe [Wed, 18 Apr 2001 03:57:05 +0000 (03:57 +0000)]
Some more tweaks to ENGINE code.

ENGINE handler functions should take the ENGINE structure as a parameter -
this is because ENGINE structures can be copied, and like other
structure/method setups in OpenSSL, it should be possible for init(),
finish(), ctrl(), etc to adjust state inside the ENGINE structures rather
than globally. This commit includes the dependant changes in the ENGINE
implementations.

23 years agoSome more tweaks to ENGINE code.
Geoff Thorpe [Wed, 18 Apr 2001 03:03:16 +0000 (03:03 +0000)]
Some more tweaks to ENGINE code.

Previous changes permanently removed the commented-out old code for where
it was possible to create and use an ENGINE statically, and this code gets
rid of the ENGINE_FLAGS_MALLOCED flag that supported the distinction with
dynamically allocated ENGINEs. It also moves the area for ENGINE_FLAGS_***
values from engine_int.h to engine.h - because it should be possible to
declare ENGINEs just from declarations in exported headers.

23 years agoSome more tweaks to ENGINE code.
Geoff Thorpe [Wed, 18 Apr 2001 02:01:36 +0000 (02:01 +0000)]
Some more tweaks to ENGINE code.

* Constify the get/set functions, and add some that functions were missing.

* Add a new 'ENGINE_cpy()' function that will produce a new ENGINE based
  copied from an original (except for the references, ie. the new copy will
  be like an ENGINE returned from 'ENGINE_new()' - a structural reference).

* Removed the "null parameter" checking in the get/set functions - it is
  legitimate to set NULL values as a way of *changing* an ENGINE (ie.
  removing a handler that previously existed). Also, passing a NULL pointer
  for an ENGINE is obviously wrong for these functions, so don't bother
  checking for it. The result is a number of error codes and strings could
  be removed.

23 years agoStructural references should never be decremented directly - so leave that
Geoff Thorpe [Wed, 18 Apr 2001 01:07:28 +0000 (01:07 +0000)]
Structural references should never be decremented directly - so leave that
to ENGINE_free(). Also, remove "#if 0" code that has no useful future.

23 years agoThis moves string constants out of vendor headers and into C files.
Geoff Thorpe [Wed, 18 Apr 2001 00:43:23 +0000 (00:43 +0000)]
This moves string constants out of vendor headers and into C files.

23 years ago'make update'
Geoff Thorpe [Tue, 17 Apr 2001 23:53:58 +0000 (23:53 +0000)]
'make update'

23 years agoClarify request of client certificates. This is a FAQ.
Lutz Jänicke [Tue, 17 Apr 2001 13:18:56 +0000 (13:18 +0000)]
Clarify request of client certificates. This is a FAQ.

23 years agoFix warning.
Ben Laurie [Mon, 16 Apr 2001 03:00:57 +0000 (03:00 +0000)]
Fix warning.

23 years agoConstify (Jason Molenda <jason@molenda.com>)
Lutz Jänicke [Sat, 14 Apr 2001 14:50:02 +0000 (14:50 +0000)]
Constify (Jason Molenda <jason@molenda.com>)

23 years agoMissing link ("Greg Stark" <gstark@ethentica.com>)
Lutz Jänicke [Thu, 12 Apr 2001 21:11:31 +0000 (21:11 +0000)]
Missing link ("Greg Stark" <gstark@ethentica.com>)

23 years agoFix wrong information with respect to CAs listed to the client
Lutz Jänicke [Thu, 12 Apr 2001 16:02:34 +0000 (16:02 +0000)]
Fix wrong information with respect to CAs listed to the client
(follows from technical discussion with Amit Chopra <amitc@pspl.co.in>).

23 years agoupdate so that changes going into the 0.9.6 tree can be logged
Bodo Möller [Thu, 12 Apr 2001 12:42:20 +0000 (12:42 +0000)]
update so that changes going into the 0.9.6 tree can be logged

23 years agoundo previous change ...
Bodo Möller [Thu, 12 Apr 2001 12:03:41 +0000 (12:03 +0000)]
undo previous change ...

23 years agoupdate from 0.9.6a
Bodo Möller [Thu, 12 Apr 2001 12:01:47 +0000 (12:01 +0000)]
update from 0.9.6a

23 years agoTypo (Jun-ichiro itojun Hagino <itojun@iijlab.net>)
Lutz Jänicke [Thu, 12 Apr 2001 11:45:42 +0000 (11:45 +0000)]
Typo (Jun-ichiro itojun Hagino <itojun@iijlab.net>)

23 years agoCorrect typo.
Richard Levitte [Wed, 11 Apr 2001 14:14:54 +0000 (14:14 +0000)]
Correct typo.

23 years agoAdd -keyform.
Richard Levitte [Wed, 11 Apr 2001 14:11:55 +0000 (14:11 +0000)]
Add -keyform.

23 years agoShow an example of moving the emailAddress object from the subkect DN
Richard Levitte [Wed, 11 Apr 2001 13:04:20 +0000 (13:04 +0000)]
Show an example of moving the emailAddress object from the subkect DN
to subjectAltName when signing a certificate.

23 years agoMake it possible to move the emailAddress object to the subjectAltName
Richard Levitte [Wed, 11 Apr 2001 12:55:06 +0000 (12:55 +0000)]
Make it possible to move the emailAddress object to the subjectAltName
extension instead of just copying it.  That makes a certificate comply
even more with PKIX recommendations according to RFC 2459.

23 years agoAdd information on 0.9.6a (in a form such that the list can be
Bodo Möller [Wed, 11 Apr 2001 10:35:38 +0000 (10:35 +0000)]
Add information on 0.9.6a (in a form such that the list can be
verified by looking at 'diff -u ../openssl-0.9.6a/CHANGES CHANGES')

23 years agoNetBSD and OpenBSD use TOD as well
Richard Levitte [Wed, 11 Apr 2001 10:06:02 +0000 (10:06 +0000)]
NetBSD and OpenBSD use TOD as well

23 years agoMention automatically queried EGD sockets (OpenSSL 0.9.7).
Bodo Möller [Tue, 10 Apr 2001 07:59:43 +0000 (07:59 +0000)]
Mention automatically queried EGD sockets (OpenSSL 0.9.7).

0.9.5 is obsolete, so we don't have to discuss its 'openssl rsa'
seeding bug.

23 years agoSome clarifications about $RANDFILE usage.
Lutz Jänicke [Mon, 9 Apr 2001 16:01:38 +0000 (16:01 +0000)]
Some clarifications about $RANDFILE usage.

23 years agoOpenSSH 1.2.2p1 is dead and gone. Errors detecting the OpenSSL library
Lutz Jänicke [Mon, 9 Apr 2001 15:55:58 +0000 (15:55 +0000)]
OpenSSH 1.2.2p1 is dead and gone. Errors detecting the OpenSSL library
are however still common and are solved by checking config.log.

23 years agoCorrect info in the FAQ.
Richard Levitte [Mon, 9 Apr 2001 14:17:26 +0000 (14:17 +0000)]
Correct info in the FAQ.

23 years agoAdjust BN_mod_inverse algorithm selection according to experiments on
Bodo Möller [Mon, 9 Apr 2001 09:28:24 +0000 (09:28 +0000)]
Adjust BN_mod_inverse algorithm selection according to experiments on
Ultra-Sparcs (both 32-bit and 64-bit compilations)

23 years agoupdate (0.9.6a)
Bodo Möller [Mon, 9 Apr 2001 07:15:16 +0000 (07:15 +0000)]
update (0.9.6a)

23 years agocomment
Bodo Möller [Sun, 8 Apr 2001 18:47:23 +0000 (18:47 +0000)]
comment

23 years agocode documentation
Bodo Möller [Sun, 8 Apr 2001 18:41:35 +0000 (18:41 +0000)]
code documentation

23 years agobinary algorithm for modular inversion
Bodo Möller [Sun, 8 Apr 2001 18:23:44 +0000 (18:23 +0000)]
binary algorithm for modular inversion

23 years agoavoid '||' since Ultrix apparently doesn't understand it
Bodo Möller [Sun, 8 Apr 2001 18:22:53 +0000 (18:22 +0000)]
avoid '||' since Ultrix apparently doesn't understand it

23 years ago'||', '&&' and 'test -x' apparently don't work on Ultrix;
Bodo Möller [Sun, 8 Apr 2001 13:49:45 +0000 (13:49 +0000)]
'||', '&&' and 'test -x' apparently don't work on Ultrix;
also 'test' appears to be available as '[' only in 'if' conditions.

23 years agoAvoid assert() in the library.
Bodo Möller [Sun, 8 Apr 2001 13:47:51 +0000 (13:47 +0000)]
Avoid assert() in the library.

23 years agoAdd forgotten "-passin" option to smime.c usage help.
Lutz Jänicke [Sun, 8 Apr 2001 10:51:14 +0000 (10:51 +0000)]
Add forgotten "-passin" option to smime.c usage help.

23 years agoResize a local buffer to accomodate the size requirements of AES.
Richard Levitte [Sun, 8 Apr 2001 05:41:42 +0000 (05:41 +0000)]
Resize a local buffer to accomodate the size requirements of AES.
Protect against future mistakes with an assert().

23 years agoAdd the possibility to have AES removed in Windows as well.
Richard Levitte [Sun, 8 Apr 2001 04:35:58 +0000 (04:35 +0000)]
Add the possibility to have AES removed in Windows as well.
Spotted by Harald Koch <chk@pobox.com>

23 years agoDon't use 'tt' uninitialized when reporting an error
Bodo Möller [Thu, 5 Apr 2001 11:40:16 +0000 (11:40 +0000)]
Don't use 'tt' uninitialized when reporting an error
(we don't have an ASN1_TEMPLATE to complain about at this stage,
so  errtt == NULL  should be OK)

23 years agoFix couple of memory leaks in PKCS7_dataDecode().
Richard Levitte [Thu, 5 Apr 2001 10:19:12 +0000 (10:19 +0000)]
Fix couple of memory leaks in PKCS7_dataDecode().
(provided by Stephen)

23 years agoUnixware config.
Richard Levitte [Thu, 5 Apr 2001 10:09:53 +0000 (10:09 +0000)]
Unixware config.

23 years agodon't use shell functions
Bodo Möller [Wed, 4 Apr 2001 16:26:31 +0000 (16:26 +0000)]
don't use shell functions

23 years agoCorrect a typo. linux != linus.
Richard Levitte [Wed, 4 Apr 2001 16:03:00 +0000 (16:03 +0000)]
Correct a typo.  linux != linus.

23 years agoIncorporate some changes that make OpenSSL compilable in CygWin.
Richard Levitte [Wed, 4 Apr 2001 15:50:30 +0000 (15:50 +0000)]
Incorporate some changes that make OpenSSL compilable in CygWin.

23 years agoSince vms.mar handles 32-bit integers, do not use it on Alpha, that's
Richard Levitte [Wed, 4 Apr 2001 13:52:56 +0000 (13:52 +0000)]
Since vms.mar handles 32-bit integers, do not use it on Alpha, that's
just a slowdown.

23 years agoOpenVMS/Alpha should use 64 bits. If nothing else, there's
Richard Levitte [Wed, 4 Apr 2001 13:51:35 +0000 (13:51 +0000)]
OpenVMS/Alpha should use 64 bits.  If nothing else, there's
performance to gain.

23 years agoMake do_bsd-gcc-shared depend on do_gnu-shared instead of the non-existent linux...
Richard Levitte [Wed, 4 Apr 2001 04:24:24 +0000 (04:24 +0000)]
Make do_bsd-gcc-shared depend on do_gnu-shared instead of the non-existent linux-shared

23 years agoFix warnings.
Bodo Möller [Tue, 3 Apr 2001 14:03:47 +0000 (14:03 +0000)]
Fix warnings.

23 years agoMake sure OPENSSL_SYS_... is defined when we need it.
Bodo Möller [Tue, 3 Apr 2001 14:03:19 +0000 (14:03 +0000)]
Make sure OPENSSL_SYS_... is defined when we need it.

23 years agoThis change should be suitable as a workaround for the Solaris x86
Bodo Möller [Tue, 3 Apr 2001 13:50:30 +0000 (13:50 +0000)]
This change should be suitable as a workaround for the Solaris x86
compiler bug reported in <01032110293775.22278@weba3.iname.net>
(the '++seq[i]' condition is evaluated as 256 rather than 0
when the previous value is 255).

23 years agoPlug a memory leak. Spotted by "Shijin" <shijin@comex.com>
Richard Levitte [Tue, 3 Apr 2001 09:42:36 +0000 (09:42 +0000)]
Plug a memory leak.  Spotted by "Shijin" <shijin@comex.com>

23 years agoReports on Windows, DG-UX and older OpenVMS.
Richard Levitte [Tue, 3 Apr 2001 09:02:07 +0000 (09:02 +0000)]
Reports on Windows, DG-UX and older OpenVMS.

23 years agolibfisdef.h and LIB do not exist on older VMS versions
Richard Levitte [Tue, 3 Apr 2001 08:31:39 +0000 (08:31 +0000)]
libfisdef.h and LIB do not exist on older VMS versions

23 years agoRemove a typo in dgux-R4-gcc.
Richard Levitte [Tue, 3 Apr 2001 08:05:03 +0000 (08:05 +0000)]
Remove a typo in dgux-R4-gcc.

23 years agoENGINE_load_[private|public]_key had error handling that could return
Geoff Thorpe [Mon, 2 Apr 2001 17:47:16 +0000 (17:47 +0000)]
ENGINE_load_[private|public]_key had error handling that could return
without releasing a lock. This is the same fix as applied to
OpenSSL-engine-0_9_6-stable, minus the ENGINE_ctrl() change - the HEAD
already had that fixed.

23 years agoActually there were two error cases that could return without releasing the
Geoff Thorpe [Mon, 2 Apr 2001 17:21:36 +0000 (17:21 +0000)]
Actually there were two error cases that could return without releasing the
lock - stupidly, my last change addressed only one of them.

23 years agoDon't return an error until the global lock is released.
Geoff Thorpe [Mon, 2 Apr 2001 17:06:36 +0000 (17:06 +0000)]
Don't return an error until the global lock is released.

23 years agoReports on VMS.
Richard Levitte [Mon, 2 Apr 2001 14:39:20 +0000 (14:39 +0000)]
Reports on VMS.

23 years agoRewrite CHOICE field setting code to properly handle
Dr. Stephen Henson [Mon, 2 Apr 2001 00:59:19 +0000 (00:59 +0000)]
Rewrite CHOICE field setting code to properly handle
combine in CHOICE options.

This was causing d2i_DSAPublicKey() to misbehave.

23 years agoA few more reports.
Richard Levitte [Sun, 1 Apr 2001 15:18:35 +0000 (15:18 +0000)]
A few more reports.

23 years agoNote reports.
Richard Levitte [Sat, 31 Mar 2001 14:03:09 +0000 (14:03 +0000)]
Note reports.

23 years agoavoid buffer overflow
Bodo Möller [Sat, 31 Mar 2001 07:48:07 +0000 (07:48 +0000)]
avoid buffer overflow

23 years agoNote the Alpha asm change
Ulf Möller [Sat, 31 Mar 2001 01:23:10 +0000 (01:23 +0000)]
Note the Alpha asm change

23 years agolinux-elf verified.
Richard Levitte [Fri, 30 Mar 2001 21:43:06 +0000 (21:43 +0000)]
linux-elf verified.

23 years agoBeta 3 has been released and announced.
Richard Levitte [Fri, 30 Mar 2001 20:54:12 +0000 (20:54 +0000)]
Beta 3 has been released and announced.

23 years agothis time *really* fix the /../ check ...
Bodo Möller [Fri, 30 Mar 2001 14:55:50 +0000 (14:55 +0000)]
this time *really* fix the /../ check ...

23 years agoFor -WWW, fix test for ".." directory references (and avoid warning for
Bodo Möller [Fri, 30 Mar 2001 10:47:21 +0000 (10:47 +0000)]
For -WWW, fix test for ".." directory references (and avoid warning for
index -1).

23 years agoNews for 0.9.6a.
Bodo Möller [Fri, 30 Mar 2001 10:46:15 +0000 (10:46 +0000)]
News for 0.9.6a.

23 years agoFix bctest, and add a workaround that should solve the problem with
Bodo Möller [Fri, 30 Mar 2001 09:19:18 +0000 (09:19 +0000)]
Fix bctest, and add a workaround that should solve the problem with
FreeBSD's /bin/sh.

23 years agoComplete the des_encrypt to des_encrypt1 rename in the main
Richard Levitte [Fri, 30 Mar 2001 07:26:54 +0000 (07:26 +0000)]
Complete the des_encrypt to des_encrypt1 rename in the main
development line as well.

23 years agoOne des_encrypt to des_encrypt1 I forgot to commit...
Richard Levitte [Thu, 29 Mar 2001 20:30:23 +0000 (20:30 +0000)]
One des_encrypt to des_encrypt1 I forgot to commit...

23 years agoStress the news about the name change.
Richard Levitte [Thu, 29 Mar 2001 15:15:10 +0000 (15:15 +0000)]
Stress the news about the name change.

23 years agoI forgot to rename des_encrypt to des_encrypt1 in libeay.num.
Richard Levitte [Thu, 29 Mar 2001 10:49:27 +0000 (10:49 +0000)]
I forgot to rename des_encrypt to des_encrypt1 in libeay.num.

23 years agoSince there has been reports of clashes between OpenSSL's
Richard Levitte [Thu, 29 Mar 2001 07:45:37 +0000 (07:45 +0000)]
Since there has been reports of clashes between OpenSSL's
des_encrypt() and des_encrypt() defined on some systems (Solaris and
Unixware and maybe others), we rename des_encrypt() to des_encrypt1().
This should have very little impact on external software unless
someone has written a mode of DES, since that's all des_encrypt() is
meant for.

23 years agoAdd news section for OpenSSL 0.9.6a. Please add what's missing
Richard Levitte [Wed, 28 Mar 2001 13:36:13 +0000 (13:36 +0000)]
Add news section for OpenSSL 0.9.6a.  Please add what's missing

23 years agocheck CRT
Ulf Möller [Wed, 28 Mar 2001 05:10:38 +0000 (05:10 +0000)]
check CRT

23 years agocheck CRT
Ulf Möller [Wed, 28 Mar 2001 04:56:58 +0000 (04:56 +0000)]
check CRT

23 years agomips
Ulf Möller [Wed, 28 Mar 2001 02:41:02 +0000 (02:41 +0000)]
mips

23 years agoRestore asm/mips3.s to be compiled and linked in, since the bug has
Richard Levitte [Tue, 27 Mar 2001 23:42:36 +0000 (23:42 +0000)]
Restore asm/mips3.s to be compiled and linked in, since the bug has
(at least hopefully :-)) been fixed.

23 years agoThe IRIX fix. Asm recap and corresponding declation.
Andy Polyakov [Tue, 27 Mar 2001 22:30:46 +0000 (22:30 +0000)]
The IRIX fix. Asm recap and corresponding declation.

Submitted by:
Reviewed by:
PR:

23 years agoInclude bn.h so we get BN_LLONG properly defined. Otherwise, we can forget things...
Richard Levitte [Tue, 27 Mar 2001 18:34:04 +0000 (18:34 +0000)]
Include bn.h so we get BN_LLONG properly defined.  Otherwise, we can forget things like %lld

23 years agoReports seem to show that asm/mips3.s has faults. To be investigated, but let's...
Richard Levitte [Tue, 27 Mar 2001 09:13:03 +0000 (09:13 +0000)]
Reports seem to show that asm/mips3.s has faults.  To be investigated, but let's avoid using it in the mean time

23 years agoFor mips3 and alpha, put the assembler file directives in separate
Richard Levitte [Mon, 26 Mar 2001 16:16:12 +0000 (16:16 +0000)]
For mips3 and alpha, put the assembler file directives in separate
variables and disable the Alpha assembler for now, since it has been
shown to fail.

The Alpha failure can be shown by adding the following numbers:

FFFFFFFFFFFFFF0000FF2E00000000EBFFFFFF0000D1
FFFFFFFFFFFFFFFFFFFFFFFFFFFFFF46FFE0FFFF0000

The result is:

1FFFFFFFFFFFEFF0000FF2E0000000032FFE0FEFF00D1

The result should really be:

1FFFFFFFFFFFFFF0000FF2E0000000032FFE0FEFF00D1

23 years agomake update
Richard Levitte [Sat, 24 Mar 2001 12:39:59 +0000 (12:39 +0000)]
make update

23 years agoWhen using the native tools on Solaris, make damn sure the native ld
Richard Levitte [Sat, 24 Mar 2001 12:31:11 +0000 (12:31 +0000)]
When using the native tools on Solaris, make damn sure the native ld
is used, even if the user has GNU ld earlier in his $PATH.

23 years agogcc uses collect2, not ld, to link things. Therefore, when using gcc
Richard Levitte [Sat, 24 Mar 2001 12:29:21 +0000 (12:29 +0000)]
gcc uses collect2, not ld, to link things.  Therefore, when using gcc
there's no need fooling ourselves, it's the gnu-shared method that we
should use.  Do it for Solaris to begin with.

23 years agoln on Solaris expects -f to come before -s.
Richard Levitte [Sat, 24 Mar 2001 11:45:57 +0000 (11:45 +0000)]
ln on Solaris expects -f to come before -s.
The linux-shared method is actually gcc-specific, so call it
gnu-shared as well.

23 years agoSmall AIX problems solved.
Richard Levitte [Thu, 22 Mar 2001 22:13:16 +0000 (22:13 +0000)]
Small AIX problems solved.

23 years agoA lot of reports.
Richard Levitte [Thu, 22 Mar 2001 21:32:26 +0000 (21:32 +0000)]
A lot of reports.

23 years agoWe really have no need for PEX_LIBS, so empty it.
Richard Levitte [Thu, 22 Mar 2001 21:20:54 +0000 (21:20 +0000)]
We really have no need for PEX_LIBS, so empty it.

23 years agoAdd missing '#ifndef OPENSSL_NO_DSA'.
Bodo Möller [Thu, 22 Mar 2001 15:06:19 +0000 (15:06 +0000)]
Add missing '#ifndef OPENSSL_NO_DSA'.