security-and-privacy/gnupg
security-and-privacy/gnupg
1
0
Fork 0
mirror of git://git.gnupg.org/gnupg.git synced 2024-12-22 10:19:57 +01:00
Code Activity

Update head to match stable 1.0

Browse Source
This commit is contained in:
David Shaw 2002-06-29 14:15:02 +00:00
parent df58e024e7
commit 0ed6228293
144 changed files with 18930 additions and 673 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

124
AUTHORS
View File

@ -1,93 +1,99 @@
Authors of GNU Privacy Guard (GnuPG)
====================================
Program: GnuPG
Maintainer: Werner Koch <wk@gnupg.org>
GNUPG Werner Koch 1998-02-23
Assigns GNU Privacy Guard and future changes.
werner.koch@guug.de
Designed and implemented GnuPG.
Authors
=======
Birger Langkjer <birger.langkjer@image.dk> Translations [da]
GNUPG Matthew Skala 1998-08-10
Disclaims changes.
mskala@ansuz.sooke.bc.ca
Wrote cipher/twofish.c.
Daniel Resare <daniel@resare.com> Translations [sv]
GNUPG Natural Resources Canada 1998-08-11
Disclaims changes by Matthew Skala.
David Shaw <dshaw@jabberwocky.com> Assignment
(all in keyserver/,
a lot of changes in g10/ see the ChangeLog,
bug fixes here and there)
Dokianakis Theofanis <madf@hellug.gr> Translations [el]
GNUPG Michael Roth Germany 1998-09-17
Assigns changes.
mroth@nessie.de
Wrote cipher/des.c.
Changes and bug fixes all over the place.
Edmund GRIMLEY EVANS <edmundo@rano.org> Translations [eo]
Gaël Quéri <gael@lautre.net> Translations [fr]
(fixed a lot of typos)
GNUPG Niklas Hernaeus 1998-09-18
Disclaims changes.
nh@df.lth.se
Weak key patches.
Gregory Steuck <steuck@iname.com> Translations [ru]
Ivo Timmermans <itimmermans@bigfoot.com> Translations [nl]
GNUPG Rémi Guyomarch 1999-05-25
Assigns past and future changes. (g10/compress.c, g10/encr-data.c,
g10/free-packet.c, g10/mdfilter.c, g10/plaintext.c, util/iobuf.c)
rguyom@mail.dotcom.fr
Jacobo Tarri'o Barreiro <jtarrio@iname.com> Translations [gl]
Janusz Aleksander Urbanowicz <alex@bofh.torun.pl> Translations [po]
TRANSLATIONS Marco d'Itri 1997-02-22
Disclaimer. [it]
md@linux.it
Magda Procha'zkova' <magda@math.muni.cz> Translations [cs]
Michael Roth <mroth@nessie.de> Assignment
(wrote cipher/des.c., changes and bug fixes all over the place)
TRANSLATIONS Gael Queri 1998-09-08
Disclaimer. [fr]
gqueri@mail.dotcom.fr
Fixed a lot of typos.
Marco d'Itri <md@linux.it> Translations [it]
Matthew Skala <mskala@ansuz.sooke.bc.ca> Disclaimer
(wrote cipher/twofish.c)
TRANSLATIONS Walter Koch 1998-09-08
Disclaimer. [de]
koch@hsp.de
Niklas Hernaeus <nh@df.lth.se> Disclaimer
(weak key patches)
Nilgun Belma Buguner <nilgun@technologist.com> Translations [tr]
TRANSLATIONS Gregory Steuck 1998-10-20
Disclaimer. [ru]
steuck@iname.com
Nils Ellmenreich <nils 'at' infosun.fmi.uni-passau.de> Assignment
(configure.in, cipher/rndlinux.c, FAQ)
Paul Eggert <eggert@twinsun.com>
(configuration macros for LFS)
TRANSLATIONS Urko Lusa
Disclaimer. [es_ES]
ulusa@euskalnet.net
Pedro Morais <morais@poli.org> Translations [pt_PT]
Rémi Guyomarch <rguyom@mail.dotcom.fr> Assignment
(g10/compress.c, g10/encr-data.c,
g10/free-packet.c, g10/mdfilter.c, g10/plaintext.c, util/iobuf.c)
TRANSLATIONS Thiago Jung Bauermann
Disclaimer. [pt_BR]
jungmann@cwb.matrix.com.br
Stefan Bellon <sbellon@sbellon.de> Assignment
(All patches to support RISC OS)
Timo Schulz <twoaday@freakmail.de> Assignment
(util/w32reg.c, g10/passphrase.c, g10/hkp.c)
TRANSLATIONS Pedro Morais
Disclaimer. [pt_PT]
morais@poli.org
Tedi Heriyanto <tedi_h@gmx.net> Translations [id]
Thiago Jung Bauermann <jungmann@cwb.matrix.com.br> Translations [pt_BR]
Rafael Caetano dos Santos <rcaetano@linux.ime.usp.br> Translations [pt_BR]
TRANSLATIONS Janusz Aleksander Urbanowicz 1999-01-09
Disclaimer. [po]
alex@bofh.torun.pl
Toomas Soome <tsoome@ut.ee> Translations [et]
Urko Lusa <ulusa@euskalnet.net> Translations [es_ES]
Walter Koch <koch@u32.de> Translations [de]
Werner Koch <wk@gnupg.org> Assignment
(started the whole thing)
Yosiaki IIDA <iida@ring.gr.jp> Translations [ja]
More credits
============
Other legal information
=======================
This program uses the zlib compression library written by
Jean-loup Gailly and Mark Adler.
Most of the stuff in mpi has been taken from the GMP library by
Torbjorn Granlund <tege@noisy.tmg.se>.
The keybox implementation is based on GDBM 1.7.3 by Philip A. Nelson.
The Rijndael implementation (cipher/rijndael.c) is based on the
public domain reference code provided for the AES selection process.
The Rijndael algorithm is due to Joan Daemen and Vincent Rijmen.
The file cipher/rndunix.c is based on rndunix.c from cryptlib.
The files cipher/rndunix.c and cipher/rndw32.c are based on rndunix.c
and rndwin32.c from cryptlib.
Copyright Peter Gutmann, Paul Kendall, and Chris Wedgwood 1996-1999.
The files in debian/ are by James Troup who is the Debian maintainer
@ -96,3 +102,13 @@ for GnuPG.
The RPM specs file scripts/gnupg.spec has been contributed by
several people.
Copyright 1998, 1999, 2000, 2001 Free Software Foundation, Inc.
This file is free software; as a special exception the author gives
unlimited permission to copy and/or distribute it, with or without
modifications, as long as this notice is preserved.
This file is distributed in the hope that it will be useful, but
WITHOUT ANY WARRANTY, to the extent permitted by law; without even the
implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.

65
BUGS
View File

@ -1,63 +1,8 @@
List of some known bugs
-------------------------
Please see
This following list contains those bugs which we are aware of. Please
make sure that bugs you report are not listed here. If you can fix one
of these bugs/limitations we will certainly be glad to receive a patch.
(Please note that we need a disclaimer if a patch is longer than about
10 lines; but it may help anyway to show us where we have to fix it. Do
an "info standards" to find out why a disclaimer is needed for GNU.)
http://www.gnupg.org/buglist.html
Format: severity: [ *] to [***], no, first reported, by, version.
Bugs which are marked with "FIX: yyyy-mm-dd" are fixed in the CVS
and after about half a day in the rsync snapshots.
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
for a list of know bugs in GnuPG. We don't distribute this list anymore
with the package because a more current one with notes in which version
the bug is fixed can be found online.
[ *] #3
--list-packets should continue even w/o a passphrase (or in batch
mode). Maybe we have to move it to a separate program??
[ *] #4 1999-01-13 <ralf.stephan@fitug.de> 0.9.1
v3 key 'expiration date' problem:
1. generate a key, set expiration date
2. <do other things, deleting etc.>
3. edit a v3 secret key, try to set expiration date
- output: "You can't change...
4. save
5. key has expiration date from 1. and gpg reports that pubkey
and seckey differ.
The for loop the exp.date is set before v3 detection?
[is this bug still alive? - can someone please check it]
[ **] #6 1999-02-22 <backes@rhrk.uni-kl.de> 0.9.3
Bus error on IRIX 6.4: Crash while doing a keygen. I think while creating
the prime. Other bus errors are reported when doing a "gpg README"
on sparc-solaris2.6.
--> Solaris fixed.
--> IRIX bug still there but someone should test this again!
(<mack@sgi.com> reports that it is still alive in 1.0.0)
(seems to be fixed now. wk 2000-01-14)
[ *] #18 1999-05-27 <Steffen.Zahn@icn.siemens.de> 0.9.7
rndunix hangs on hp/ux. The problem is related to my_plcose which is
not always called. (I suggest to use EGD instead of rndunix.)
[ *] #24 1999-09-05 <marcus@yoyo.cc.monash.edu.au> 0.9.11
Does not link libc and libz expect when removing all "-lc -lz" except
for the last one. This happens on some systems when not using the
GNU ld. This need to be fixed in libtool.
[ *] #28 1999-11-29
--list-key will only emit an error about unknown UIDs when all UIDs
are unknown.
[Postponed for 1.2]
[ *] #30 1999-12-16 <pfeifer@dbai.tuwien.ac.at>
Using -r with a name does not check for multiple recpients and may
lead to unwanted recipients.
[ *] #31 199-12-24 Jeff Allen <jra@corp.webtv.net>
VPATH build in a subdirectory of the sources does not work for po files.
Next #31

14
COPYING
View File

@ -11,7 +11,7 @@
The licenses for most software are designed to take away your
freedom to share and change it. By contrast, the GNU General Public
License is intended to guarantee your freedom to share and change free
software--to make sure the software is free for all its users. This
software--to make sure the software is free for all its users. This
General Public License applies to most of the Free Software
Foundation's software and to any other program whose authors commit to
using it. (Some other Free Software Foundation software is covered by
@ -19,7 +19,7 @@ the GNU Library General Public License instead.) You can apply it to
your programs, too.
When we speak of free software, we are referring to freedom, not
price. Our General Public Licenses are designed to make sure that you
price. Our General Public Licenses are designed to make sure that you
have the freedom to distribute copies of free software (and charge for
this service if you wish), that you receive source code or can get it
if you want it, that you can change the software or use pieces of it
@ -201,7 +201,7 @@ otherwise) that contradict the conditions of this License, they do not
excuse you from the conditions of this License. If you cannot
distribute so as to satisfy simultaneously your obligations under this
License and any other pertinent obligations, then as a consequence you
may not distribute the Program at all. For example, if a patent
may not distribute the Program at all. For example, if a patent
license would not permit royalty-free redistribution of the Program by
all those who receive copies directly or indirectly through you, then
the only way you could satisfy both it and this License would be to
@ -239,7 +239,7 @@ of the General Public License from time to time. Such new versions will
be similar in spirit to the present version, but may differ in detail to
address new problems or concerns.
Each version is given a distinguishing version number. If the Program
Each version is given a distinguishing version number. If the Program
specifies a version number of this License which applies to it and "any
later version", you have the option of following the terms and conditions
either of that version or of any later version published by the Free
@ -249,7 +249,7 @@ Foundation.
10. If you wish to incorporate parts of the Program into other free
programs whose distribution conditions are different, write to the author
to ask for permission. For software which is copyrighted by the Free
to ask for permission. For software which is copyrighted by the Free
Software Foundation, write to the Free Software Foundation; we sometimes
make exceptions for this. Our decision will be guided by the two goals
of preserving the free status of all derivatives of our free software and
@ -291,7 +291,7 @@ convey the exclusion of warranty; and each file should have at least
the "copyright" line and a pointer to where the full notice is found.
<one line to give the program's name and a brief idea of what it does.>
Copyright (C) 19yy <name of author>
Copyright (C) <year> <name of author>
This program is free software; you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
@ -313,7 +313,7 @@ Also add information on how to contact you by electronic and paper mail.
If the program is interactive, make it output a short notice like this
when it starts in an interactive mode:
Gnomovision version 69, Copyright (C) 19yy name of author
Gnomovision version 69, Copyright (C) year name of author
Gnomovision comes with ABSOLUTELY NO WARRANTY; for details type `show w'.
This is free software, and you are welcome to redistribute it
under certain conditions; type `show c' for details.

515
ChangeLog
View File

@ -1,151 +1,486 @@
2000-12-19 Werner Koch <wk@gnupg.org>
2002-06-20 David Shaw <dshaw@jabberwocky.com>
Removed the stuff which is now in libgcrypt. These are
mainly the directories mpi/, cipher/ and gcrypt/.
* NEWS: changes since 1.0.7.
2000-11-14 Werner Koch <wk@gnupg.org>
* configure.ac: Set new version number (1.1.90), and fix Solaris
compiler flags for shared objects.
Version 1.1.2 released.
2002-06-11 David Shaw <dshaw@jabberwocky.com>
2000-11-13 Werner Koch <wk@gnupg.org>
* configure.ac: Move -lsocket and -lnsl checks before LDAP link
tests so they work properly on Solaris. Noted by David Champion.
Also, check for the Mozilla LDAP library if the OpenLDAP library
check fails. Put -lsocket and -lnsl in NETLIBS rather than LIBS
so not all programs are forced to link to them.
2002-06-05 David Shaw <dshaw@jabberwocky.com>
* acinclude.m4 (GNUPG_FIX_HDR_VERSION): VPATH build fix.
* configure.ac: Add a switch for the experimental external HKP
keyserver interface.
2000-10-10 Werner Koch <wk@gnupg.org>
2002-05-22 Werner Koch <wk@gnupg.org>
* Makefile.am (dist-hook): Create the version file.
* configure.in: Set the libtool version here, removed the need
for the version file.
* configure.ac: Check for strcasecmp and strncasecmp. Removed
stricmp and memicmp checks.
Mon Sep 18 16:35:45 CEST 2000 Werner Koch <wk@openit.de>
2002-05-08 David Shaw <dshaw@jabberwocky.com>
* acinclude.m4 (GNUPG_CHECK_MLOCK): Removed that silly mkdir().
* configure.ac: If LDAP comes up unusable, try #including <lber.h>
before giving up. Old versions of OpenLDAP require that.
* configure.in: Changes to allow for Solaris random device.
By Nils Ellmenreich.
(--with-egd-socket): New.
2002-05-03 David Shaw <dshaw@jabberwocky.com>
* configure.in (GNUPG_HOMEDIR): New.
* configure.ac: In g10defs.h, use \ for the directory separator
when HAVE_DOSISH_SYSTEM is on.
* configure.in: Check for fstat64 and fopen64
* configure.ac: Add --disable-exec flag to disable all remote
program execution. --disable-exec implies --disable-ldap and
--disable-mailto. Also look in /usr/lib for sendmail. If
sendmail is not found, do not default - just fail.
2002-04-30 David Shaw <dshaw@jabberwocky.com>
* configure.ac: Try and link to a sample LDAP program to check if
the LDAP we're about to use is really sane. The most common
problem (using a very old OpenLDAP), could be fixed with an extra
#include, but this would not be very portable to other LDAP
libraries.
2002-04-29 Werner Koch <wk@gnupg.org>
Released 1.0.7.
* README: Fixed some minor things.
2002-04-25 Werner Koch <wk@gnupg.org>
* configure.ac: Check for locale.h and setlocale
2002-04-24 David Shaw <dshaw@jabberwocky.com>
* Update NEWS with recent changes.
2002-04-19 Werner Koch <wk@gnupg.org>
Released 1.0.6e snapshot.
2002-04-12 Werner Koch <wk@gnupg.org>
* configure.ac: Add a warning note to the definition of the
EXTSEP macros.
2002-04-09 Werner Koch <wk@gnupg.org>
* configure.ac (ALL_LINGUAS): Added Czech, Galician and Greek
translations. s/es_ES/es/.
2002-03-06 Werner Koch <wk@gnupg.org>
* configure.ac (ALL_LINGUAS): s/pt_PT/pt/
2002-03-04 David Shaw <dshaw@jabberwocky.com>
* Add a AC_DEFINE(_GNU_SOURCE). Since this is always defined in
config.h, the various autoconf tests should be tested with it
enabled. This also works around a compiler warning caused by a
minor header bug in glibc 2.1 that causes fseeko to be defined
when building gpg, but not when tested for in configure.
2002-03-03 Werner Koch <wk@gnupg.org>
Release 1.0.6d snapshot.
2002-01-04 David Shaw <dshaw@jabberwocky.com>
* NEWS: about symmetric messages and fixed file sizes.
2001-12-22 Werner Koch <wk@gnupg.org>
Released 1.0.6c snapshot.
* configure.ac (AH_BOTTOM): Moved EXEC_TEMPFILE_ONLY to here.
* acconfig.h: Removed, it should no longer be used.
2001-12-21 David Shaw <dshaw@jabberwocky.com>
* Add an acconfig.h to define EXEC_TEMPFILE_ONLY on platforms that
can't do fork/exec.
2001-12-21 Werner Koch <wk@gnupg.org>
* Makefile.am (dist-hook): We should also look in include for
distfiles.
(EXTRA_DIST): Remove VERSION because it is generated by dist-hook.
2001-12-20 David Shaw <dshaw@jabberwocky.com>
* configure.ac: replacement function for mkdtemp()
2001-12-19 David Shaw <dshaw@jabberwocky.com>
* configure.ac: Check for stat()
2001-12-19 Werner Koch <wk@gnupg.org>
* acinclude.m4 (GNUPG_SYS_SYMBOL_UNDERSCORE): Support Cygwin target
* configure.ac [CYGWIN32]: Don't build tiger. By Disastry.
2001-12-18 Werner Koch <wk@gnupg.org>
* Makefile.am (SUBDIRS): Add keyserver.
(dist-hook): Only look in mpi and scripts for distfiles; this way
we don't include those of a stale "make dist" directory.
2001-10-23 Werner Koch <wk@gnupg.org>
Released 1.0.6b snapshot.
2001-10-22 Werner Koch <wk@gnupg.org>
* configure.ac: Fixed for automake 1.5
2001-10-17 Werner Koch <wk@gnupg.org>
* README: Removed note on local_ID.
2001-09-28 Werner Koch <wk@gnupg.org>
* configure.ac: From now on add a string "-cvs" to the version to
indicate that this is a pre-release of the given version number.
2001-09-26 Werner Koch <wk@gnupg.org>
* configure.ac [MINGW32]: Switched from wsock.dll to ws2_32.dll.
2001-09-09 Werner Koch <wk@gnupg.org>
* configure.ac: autoconf changed the name of the maintainer mode
flag, so that not all warnings where enabled. Fixed that.
Reported by Dirk Meyer.
2001-09-07 Werner Koch <wk@gnupg.org>
* configure.ac: Test for strsep().
2001-09-03 Werner Koch <wk@gnupg.org>
* configure.ac: Removed GDBM tests.
2001-08-23 Werner Koch <wk@gnupg.org>
* configure.in (AC_FUNC_FSEEKO): Add.
(AC_CHECK_FUNCS): Remove fseeko. By Paul Eggert <eggert@twinsun.com>.
2001-08-22 Werner Koch <wk@gnupg.org>
* configure.ac (gethrtime): Enhanced the test by running a test
program.
* INSTALL: Removed the note about Solaris problems because the
above test should catch this.
2001-08-20 Werner Koch <wk@gnupg.org>
* acinclude.m4: Add check for plock if mlock is broken.
* configure.ac: Use regular tests for -lsocket and -lnsl,
more thorough test for gethrtime, allow specifying the path to
the zlib library if it is not in the default compiler/linker
search path, use ${datadir}. All these test enhancements are by
Albert Chin.
* configure.ac: Set some compiler flags for dec-osf and hpux. By
Tim Mooney.
* configure.ac: Create g10defs.h with EXTSEP_S et al.
2001-08-03 Werner Koch <wk@gnupg.org>
* configure.ac (VERSION,PACKAGE): Fixed quoting.
2001-07-26 Werner Koch <wk@gnupg.org>
* configure.ac: Finally got it running with the new autoconf. Had
to define PACKAGE and VERSION and to add -I.. to each Makefile.am.
2001-07-09 Werner Koch <wk@gnupg.org>
Migrated to autoconf 2.50.
* acinclude.m4: Removed the temporary LFS macros and GNUPG_LINK_FILES.
* acconfig.h: Removed
* configure.in: Replaced by...
* configure.ac: and modified for use with autoconf 2.50, use a
literal string for the version number. Replaced GNUPG_LINK_FILES
with AC_CONFIG_LINKS and moved some informational messages to the end.
* VERSION: Removed.
* Makefile.am (DISTCLEANFILES): gettext is better now; no more
need to remove the libintl.h symlink.
(dist-hook): Create VERSION file.
2001-06-08 Werner Koch <wk@gnupg.org>
* configure.in (DYNLINK_MOD_CFLAGS): Use -shared with dec-osf.
Reported by Chris Adams. Merged some cases.
2001-05-29 Werner Koch <wk@gnupg.org>
Released version 1.0.6.
2001-05-28 Werner Koch <wk@gnupg.org>
* configure.in (BUILD_INCLUDED_LIBINTL): Set to no for W32.
2001-04-29 Werner Koch <wk@gnupg.org>
Released version 1.0.5.
2001-04-28 Werner Koch <wk@gnupg.org>
Updated all copyright notices.
2001-04-27 Werner Koch <wk@gnupg.org>
* README: Removed a few outdated paragraphs.
2001-04-17 Werner Koch <wk@gnupg.org>
* acinclude.m4 (GNUPG_CHECK_DOCBOOK_TO_TEXI): New.
(GNUPG_CHECK_TYPEDEF): Define _GNU_SOURCE because we will use it
anyway.
* configure.in: Use it here.
* configure.in (ALL_LINGUAS): Add Estonian translation by Toomas Soome.
(use_m_debug): Removed --enable-m-debug because it does not work
anymore.
2001-04-06 Werner Koch <wk@gnupg.org>
* configure.in (ALL_LINGUAS): Add Turkish translation. Thanks
to Nilgun Belma Buguner.
2001-03-18 Werner Koch <wk@gnupg.org>
* configure.in: Hardwire the use of -lsocket for some
systems. Thanks to Reinhard Wobst.
2001-03-13 Werner Koch <wk@gnupg.org>
* configure.in: Add copyright notice and -lwsock32.
2001-03-12 Werner Koch <wk@gnupg.org>
* INSTALL: Add a note to VPATH builds.
2001-03-08 Werner Koch <wk@gnupg.org>
* debian/: Applied update from James.
Added copyright and license notices to some more files.
2001-01-18 Werner Koch <wk@gnupg.org>
* configure.in: Removed tool definitions for MingW32
2000-11-17 Werner Koch <wk@gnupg.org>
* acinclude.m4 (GNUPG_CHECK_FAQPROG): Do not prinnt the warning.
2000-11-11 Paul Eggert <eggert@twinsun.com>
Actually Paul's patches are dated 2000-20-17; I applied them today
and merged some ChangeLog entries (wk@gnupg.org).
* configure.in (AC_SYS_LARGEFILE): Add.
(try_large_file): Remove. All uses removed.
(AC_CHECK_FUNCS): Remove fopen64 and fstat64.
* acinclude.m4 (AC_SYS_LARGEFILE_TEST_INCLUDES,
AC_SYS_LARGEFILE_MACRO_VALUE, AC_SYS_LARGEFILE): New macros,
taken from GNU tar.
* configure.in (AC_CHECK_FUNCS): Add fseeko.
2000-10-17 Werner Koch <wk@gnupg.org>
* configure.in: Disabled fopen64 checks for Solaris and HPUX.
2000-10-13 Werner Koch <wk@gnupg.org>
* configure.in: Append mpi/mpi-asm-defs.h to g10defs.h
2000-10-09 Werner Koch <wk@gnupg.org>
* acinclude.m4: Changed wording of the faqprog.pl warning.
Wed Oct 4 15:50:18 CEST 2000 Werner Koch <wk@openit.de>
* configure.in: Set DYNLINK_MOD_CFLAGS for Irix. It seems that Irix
needs the -shared flag. In 1.1 we are going to use libtool, so this
module stuff will get redesigned anyway. Suggested by Jeff Long.
Thu Sep 14 14:20:38 CEST 2000 Werner Koch <wk@openit.de>
* acinclude.m4 (GNUPG_CHECK_FAQPROG): New.
* configure.in: Test for this.
* configure.in (DYNLINK_MOD_CFLAGS): Fix by David Champion.
Tue Aug 22 14:31:15 CEST 2000 Werner Koch <wk@openit.de>
Wed Sep 6 17:55:47 CEST 2000 Werner Koch <wk@openit.de>
Version 1.1.1
* configure.in: Check for fstat64 and fopen64
Fri Aug 18 14:27:14 CEST 2000 Werner Koch <wk@openit.de>
Wed Sep 6 14:59:09 CEST 2000 Werner Koch <wk@openit.de>
* agent/: New.
* Makefile.am, configure.in: Support for the new directory.
* configure.in (GNUPG_HOMEDIR): New.
Mon Jul 17 16:35:47 CEST 2000 Werner Koch <wk@>
Fri Aug 25 16:05:38 CEST 2000 Werner Koch <wk@openit.de>
* configure.in: Changes to allow for Solaris random device.
By Nils Ellmenreich.
(--with-egd-socket): New.
Wed Aug 23 19:52:51 CEST 2000 Werner Koch <wk@openit.de>
* acinclude.m4 (GNUPG_CHECK_MLOCK): Removed that silly mkdir().
Wed Jul 19 11:26:43 CEST 2000 Werner Koch <wk@openit.de>
* configure.in (mingw32): Changes to allow for mingw32msvc
Fri Jul 14 19:38:23 CEST 2000 Werner Koch <wk@>
The big merge between this one and the stable branch 1.0. Still need
to merge TNANKS, AUTHORS and such. It probaly does not compile yet.
Fri Jul 14 10:17:30 CEST 2000 Werner Koch <wk@openit.de>
* acinclude.m4 (GNUPG_CHECK_MLOCK): Fixed syntax error in C code.
Wed Jul 12 13:32:06 CEST 2000 Werner Koch <wk@openit.de>
Version 1.0.2
Fri Jun 9 10:09:52 CEST 2000 Werner Koch <wk@openit.de>
* configure.in: Add check for termio.h, wait unctiosn and sigaction.
* acinclude.m4, configure.in (GNUPG_CHECK_GNUMAKE): New.
Wed Jun 7 19:19:09 CEST 2000 Werner Koch <wk@openit.de>
* acinclude.m4 (MKDIR_TAKES_ONE_ARG): Check some headers. By Gaël Quéri.
* configure.in (AM_INIT_AUTOMAKE): Use this now. By Gaël.
* acinclude.m4 (GNUPG_CHECK_EXPORTDYNAMIC): Replacement for
Mon Jun 5 12:37:43 CEST 2000 Werner Koch <wk@openit.de>
* acnclude.m4 (GNUPG_CHECK_EXPORTDYNAMIC): Replacement for
GNUPG_CHECK_RDYNAMIC which should handle gcc with non GNU ld nicer.
Contributed by Dave Dykstra.
* configure.in (GNYPG_CHECK_RDYNAMIC): Replaced by the new check.
* configure.in: Add a test for unisgned long long.
* configure.in (AC_CHECK_AWK): Moved before the first use of AWK. Suggested
by Dave Dykstra.
* configure.in (DYNLINK_MOD_CFLAGS): Set different for NetBSD.
Tue May 30 16:37:55 CEST 2000 Werner Koch <wk@openit.de>
* configure.in: Add check for clock_gettime
Version 1.0.1-ePit-1
* configure.in (ALL_LINGUAS): Add nl.
* configure.in (ALL_LINGUAS): Add Esperanto.
* configure.in (ALL_LINGUAS): Add sv and ja.
Sun May 28 13:55:17 CEST 2000 Werner Koch <wk@openit.de>
* configure.in: Use /usr/local for CFLAGS and LDFLAGS when
target is freebsd. By Rémi.
* acinclude.m4 (GNUPG_SYS_NM_PARSE): Added BSDI support.
(GNUPG_CHECK_RDYNAMIC): Ditto.
* configure.in: Do not set development version when the version has
a dash in it. Suggested by Dave Dykstra.
Wed Apr 19 10:57:26 CEST 2000 Werner Koch <wk@openit.de>
* configure.in: Removed substitution for doc/gph/Makefile.
Do all the gcc warning only in maintainer mode.
* acconfig.h (HAVE_MLOCK): Added
* configure.in (dlopen): Use CHECK_FUNC for a test of dlopen in libc.
Suggested by Alexandre Oliva.
(-Wall): Moved the settting of gcc warning options near to the end
so that tests don't get confused. Suggested by Paul D. Smith.
Wed Mar 22 13:50:24 CET 2000 Werner Koch <wk@openit.de>
* acinclude.m4 (GNUPG_SYS_NM_PARSE): Added BSDI support.
(GNUPG_CHECK_RDYNAMIC): Ditto.
* acinclude.m4 (GNUPG_CHECK_MLOCK): Changed the way to test for
librt. Test suggested by Jeff Long.
* acinclude.m4 (GNUPG_CHECK_MLOCK): Changed the way to test for
librt. Test suggested by Jeff Long.
Fri Mar 17 17:50:25 CET 2000 Werner Koch <wk@openit.de>
* acinclude.m4 (GNUPG_CHECK_MLOCK): Do librt check only when
we can't link a test program. This way GNU systems don't need
to link against linrt.
(GNUPG_CHECK_IPC): Fixed use of TRY_COMPILE macro. From Tim Mooney.
* acinclude.m4 (GNUPG_CHECK_MLOCK): Do librt check only when
we can't link a test program. This way GNU systems don't need
to link against linrt.
(GNUPG_CHECK_IPC): Fixed use of TRY_COMPILE macro. From Tim Mooney.
* acinclude.m4 (GNUPG_SYS_SYMBOL_UNDERSCORE): Add support for
DJGPP.
(GNUPG_CHECK_MLOCK): Check whether mlock sits in librt.
2000-03-14 12:07:54 Werner Koch (wk@habibti.openit.de)
* acinclude.m4 (GNUPG_CHECK_RDYNAMIC): Add NetBSD. By Thomas Klausner.
* acinclude.m4 (GNUPG_SYS_SYMBOL_UNDERSCORE): Add support for
DJGPP.
(GNUPG_CHECK_MLOCK): Check whether mlock sits in librt.
* configure.in: Add a test for unisgned long long.
* acconfig.h (HAVE_MLOCK): Added
Tue Mar 7 18:45:31 CET 2000 Werner Koch <wk@gnupg.de>
Mon Mar 13 19:22:46 CET 2000 Werner Koch <wk@openit.de>
* acinclude.m4 (GNUPG_CHECK_RDYNAMIC): Add NetBSD. By Thomas Klausner.
* configure.in (DYNLINK_MOD_CFLAGS): Set different for NetBSD.
* configure.in: Now uses the Docbook M4s from GPH.
Thu Mar 2 15:37:46 CET 2000 Werner Koch <wk@gnupg.de>
Mon Jan 31 17:46:35 CET 2000 Werner Koch <wk@>
* configure.in: Add check for clock_gettime
* Makefile.am: Re-added tools. By Rémi.
Wed Feb 23 10:07:57 CET 2000 Werner Koch <wk@gnupg.de>
Mon Jan 31 16:37:34 CET 2000 Werner Koch <wk@gnupg.de>
* configure.in (ALL_LINGUAS): Add nl.
* configure.in: Create a symlink for types.h in gcrypt/.
Wed Feb 16 16:25:09 CET 2000 Werner Koch <wk@gnupg.de>
Thu Jan 27 18:00:44 CET 2000 Werner Koch <wk@gnupg.de>
* configure.in (ALL_LINGUAS): Add Esperanto.
* configure.in (g10defs.h): Replaced by gnupg-defs.h
Wed Feb 16 14:09:00 CET 2000 Werner Koch <wk@gnupg.de>
Mon Jan 24 13:04:28 CET 2000 Werner Koch <wk@gnupg.de>
* configure.in (ALL_LINGUAS): Add sv and ja.
* jnlib/ : New.
* AUTHORS: Converted to a more compact format.
* configure.in: Do set development version when the version has
* INSTALL: Wrote a note about a Solaris problem.
Thu Feb 10 17:39:44 CET 2000 Werner Koch <wk@gnupg.de>
* configure.in: Use /usr/local for CFLAGS and LDFLAGS when
target is freebsd. By Rémi.
Thu Jan 13 19:31:58 CET 2000 Werner Koch <wk@gnupg.de>
* configure.in: Do not set development version when the version has
a dash in it. Suggested by Dave Dykstra.
Thu Dec 9 17:22:27 CET 1999 Werner Koch <wk@gnupg.de>
Thu Dec 16 10:07:58 CET 1999 Werner Koch <wk@gnupg.de>
* acinclude.m4 (GNUPG_FIX_HDR_VERSION): New.
* configure.in: Check and fix the version number of gcrypt/gcrypt.h
so that it is always the save as VERSION.
* VERSION: Set to 1.0.1.
Thu Oct 28 16:17:46 CEST 1999 Werner Koch <wk@gnupg.de>
* configure.in: Removed substitution for doc/gph/Makefile.
Do all the gcc warning only in maintainer mode.
* Started with development series 1.1 on 1999-10-26
Thu Dec 9 10:31:05 CET 1999 Werner Koch <wk@gnupg.de>
Tue Oct 26 14:10:21 CEST 1999 Werner Koch <wk@gnupg.de>
* INSTALL: Add a hint for AIX. By Jos Backus.
* README-alpha: New
Sat Dec 4 12:30:28 CET 1999 Werner Koch <wk@gnupg.de>
* configure.in (dlopen): Use CHECK_FUNC for a test of dlopen in libc.
Suggested by Alexandre Oliva.
(-Wall): Moved the settting of gcc warning options near to the end
so that tests don't get confused. Suggested by Paul D. Smith.
Mon Nov 22 11:14:53 CET 1999 Werner Koch <wk@gnupg.de>
* BUGS: Replaced content with a link to the online list.
Fri Nov 12 20:33:19 CET 1999 Werner Koch <wk@gnupg.de>
* README: Fixed a type and add a note about the gnupg-i18n ML.
Thu Oct 28 16:08:20 CEST 1999 Werner Koch <wk@gnupg.de>
* acinclude.m4, configure.in (GNUPG_CHECK_GNUMAKE): New.
Sat Oct 9 20:34:41 CEST 1999 Werner Koch <wk@gnupg.de>
* configure.in: Tweaked handling of random modules and removed
dummy support for libgcrypt.
* Makefile.am: Removed libgcrypt support.
* cgrypt/ : Removed.
* Makefile.am: Removed libtool.
Fri Oct 8 20:32:01 CEST 1999 Werner Koch <wk@gnupg.de>
* configure.in: Fixed quoting in test for development version.
@ -153,17 +488,18 @@ Tue Oct 26 14:10:21 CEST 1999 Werner Koch <wk@gnupg.de>
very nice Howto documents - I apoligize for forgetting to mention them
earlier.
Fri Sep 17 12:56:42 CEST 1999 Werner Koch <wk@isil.d.shuttle.de>
Tue Sep 28 20:54:37 CEST 1999 Werner Koch <wk@gnupg.de>
* textfilter.c (copy_clearsig_text) [__MINGW32__): Use CR,LF.
Fri Sep 17 12:56:42 CEST 1999 Werner Koch <wk@isil.d.shuttle.de>
* configure.in: Add "-lcap" when capabilities are requested.
Add the conditional CROSS_COMPILING.
* Makefile.am: Don't use checks when CROSS_COMPILING.
Wed Sep 15 16:22:17 CEST 1999 Werner Koch <wk@isil.d.shuttle.de>
* configure.in (ALL_LINGUAS): Add pt_PT.
* configure.in: Some tweaks for cross compiling under MingW32
@ -171,12 +507,10 @@ Wed Sep 15 16:22:17 CEST 1999 Werner Koch <wk@isil.d.shuttle.de>
Tue Sep 7 17:08:10 CEST 1999 Werner Koch <wk@isil.d.shuttle.de>
* VERSION: Set to 1.0.0.
Mon Sep 6 19:59:08 CEST 1999 Werner Koch <wk@isil.d.shuttle.de>
* configure.in: Create makefile in doc/gph
* acinclude.m4 (GNUPG_FUNC_MKDIR_TAKES_ONE_ARG): New
@ -184,7 +518,6 @@ Mon Sep 6 19:59:08 CEST 1999 Werner Koch <wk@isil.d.shuttle.de>
Thu Sep 2 16:40:55 CEST 1999 Werner Koch <wk@isil.d.shuttle.de>
* VERSION: Set to 0.9.11.
Tue Aug 31 17:20:44 CEST 1999 Werner Koch <wk@isil.d.shuttle.de>
@ -529,3 +862,15 @@ Fri Feb 13 19:43:41 1998 Werner Koch (wk@isil.d.shuttle.de)
* configure.in : Fixed zlib stuff
* Makefile.am: Likewise
Copyright 1998, 1999, 2000, 2001, 2002 Free Software Foundation, Inc.
This file is free software; as a special exception the author gives
unlimited permission to copy and/or distribute it, with or without
modifications, as long as this notice is preserved.
This file is distributed in the hope that it will be useful, but
WITHOUT ANY WARRANTY, to the extent permitted by law; without even the
implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.

66
INSTALL
View File

@ -1,3 +1,14 @@
Installation instructions for GnuPG
====================================
Copyright 1998, 1999, 2000, 2001 Free Software Foundation, Inc.
This file is free software; as a special exception the author gives
unlimited permission to copy and/or distribute it, with or without
modifications, as long as this notice is preserved.
This file is distributed in the hope that it will be useful, but
WITHOUT ANY WARRANTY, to the extent permitted by law; without even the
implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
Please read the Basic Installation section somewhere below.
@ -18,6 +29,15 @@ Configure options for GNUPG
none - Do not linkl any module in but rely on
a dynmically loaded modules.
--with-egd-socket=<name> This is only used when EGD is used as random
gatherer. GnuPG uses by default "~/.gnupg/entropy"
as the socket to connect EGD. Using this option the
socket name can be changed. You may use any filename
here with 2 exceptions: a filename starting with
"~/" uses the socket in the homedirectory of the user
and one starting with a "=" uses a socket in the
GnuPG homedirectory which is bye default "~/.gnupg".
--with-included-zlib Forces usage of the local zlib sources. Default is
to use the (shared) library of the system.
@ -26,7 +46,10 @@ Configure options for GNUPG
--disable-nls Disable NLS support (See the file ABOUT-NLS)
--enable-m-guard Enable the integrated malloc checking code.
--enable-m-guard Enable the integrated malloc checking code. Please
note that this feature does not work on all CPUs
(e.g. SunOS 5.7 on UltraSparc-2) and might give
you a Bus error.
--disable-dynload If you have problems with dynamic loading, this option
disables all dynamic loading stuff.
@ -70,6 +93,28 @@ This doesn't matter and we know about it (actually it is due to the some
warning options which we have enabled for gcc)
Specific problems on some machines
==================================
* IBM RS/6000 running AIX:
Due to a change in gcc (since version 2.8) the MPI stuff may
not build. In this case try to run configure using:
CFLAGS="-g -O2 -mcpu=powerpc" ./configure
* Compaq C V6.2 for alpha:
You may want to use the option "-msg-disable ptrmismatch1"
to get rid of the sign/unsigned char mismatch warnings.
* SVR4.2 (ESIX V4.2 cc)
Due to problems with the ESIX as, you probably want to do
CFLAGS="-O -K pentium" ./configure --disable-asm
Reported by Reinhard Wobst.
The Random Device
=================
Random devices are available in Linux, FreeBSD and OpenBSD.
@ -112,6 +157,10 @@ We will use capabilities in the future.
The ~/.gnupg directory will be created if it does not exist. Your first
action should be to create a key pair: "gpg --gen-key".
Due to limitations in the automake system, the Info format versions of
the man pages are not installed. You have to convert the Texinfo
files by hand (use makeinfo) and copy them to the appropriate place.
Creating a RPM package
@ -201,13 +250,14 @@ Or, on systems that have the `env' program, you can do it like this:
Compiling For Multiple Architectures
====================================
You can compile the package for more than one kind of computer at the same
time by placing the object files for each architecture in their own
directory. To do this, you must use a version of `make', such as GNU `make',
that supports the `VPATH' variable. `cd' to the directory where you want the
object files and executables to go and run the `configure' script.
`configure' automatically checks for the source code in the directory that
`configure' is in and in `..'.
You can compile the package for more than one kind of computer at
the same time by placing the object files for each architecture in
their own directory. To do this, you must use a version of `make',
such as GNU `make', that supports the `VPATH' variable. `cd' to the
directory where you want the object files and executables to go and
run the `configure' script; please use a relative filename name to
invoke `configure'. `configure' automatically checks for the source
code in the directory that `configure' is in and in `..'.
If you have to use a `make' that does not supports the `VPATH'
variable, you have to compile the package for one architecture at a time

45
Makefile.am
View File

@ -1,3 +1,21 @@
# Copyright (C) 1998, 1999, 2000, 2001 Free Software Foundation, Inc.
#
# This file is part of GnuPG.
#
# GnuPG is free software; you can redistribute it and/or modify
# it under the terms of the GNU General Public License as published by
# the Free Software Foundation; either version 2 of the License, or
# (at your option) any later version.
#
# GnuPG is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with this program; if not, write to the Free Software
# Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA
## Process this file with automake to produce Makefile.in
if CROSS_COMPILING
@ -6,27 +24,17 @@ else
checks = checks
endif
if COMPILE_AGENT
agent = agent
else
agent =
endif
SUBDIRS = intl zlib jnlib util g10 po tools \
${agent} doc ${checks}
EXTRA_DIST = README-alpha PROJECTS BUGS
# gettext never gets it right, so we take here care of deleting the
# symlink. my_clean_gcrypt is just a kludge until we can include
# libgcrypt.
DISTCLEANFILES = gnupg-defs.h intl/libintl.h
SUBDIRS = intl zlib util mpi cipher tools g10 keyserver po doc ${checks}
EXTRA_DIST = PROJECTS BUGS config.h.in
DISTCLEANFILES = g10defs.h
# Add all the files listed in "distfiles" files to the distribution,
# apply version number s to some files and create a VERSION file which
# apply version numbers to some files and create a VERSION file which
# we need for the Prereq: patch file trick.
dist-hook:
@set -e; \
for file in `cd $(top_srcdir); find . -type f -name distfiles`; do \
for file in `cd $(top_srcdir); \
find scripts mpi include -type f -name distfiles`; do \
dir=`dirname $$file` ; $(mkinstalldirs) $(distdir)/$$dir ; \
for i in distfiles `cat $(top_srcdir)/$$file` ; do \
ln $(top_srcdir)/$$dir/$$i $(distdir)/$$dir/$$i 2> /dev/null \
@ -36,8 +44,11 @@ dist-hook:
@set -e; \
sed -e 's/@pkg_version@/$(VERSION)/g' \
$(top_srcdir)/scripts/gnupg.spec.in \
> $(distdir)/scripts/gnupg.spec ; \
> $(distdir)/scripts/gnupg.spec
echo "$(VERSION)" > $(distdir)/VERSION

304
NEWS
View File

@ -1,15 +1,258 @@
Noteworthy changes in version 1.1.90
------------------------------------------------
GnuPG now needs libgcrypt - you will find it at the same place
where you got GnuPG.
* New commands: --personal-cipher-preferences,
--personal-digest-preferences, and
--personal-compress-preferences allow the user to specify which
algorithms are to be preferred. Note that this does not permit
using an algorithm that is not present in the recipient's
preferences (which would violate the OpenPGP standard). This
just allows sorting the preferences differently.
Noteworthy changes in version 1.1.2
-----------------------------------
* New "group" command to refer to several keys with one name.
--> THIS IS A DEVELOPMENT VERSION; see README and README-alpha <--
* A warning is issued if the user forces the use of an algorithm
that is not listed in the recipient's preferences.
* Add Rijndael (AES) support.
* Full revocation key (aka "designated revoker") support.
* Removed gdbm support.
* The preferred hash algorithms on a key are consulted when
encrypting a signed message to that key. Note that this is
disabled by default by a SHA1 preference in
--personal-digest-preferences.
* --cert-digest-algo allows the user to specify the hash algorithm
to use when signing a key rather than the default SHA1 (or MD5
for PGP2 keys). Do not use this feature unless you fully
understand the implications of this.
* --pgp7 mode automatically sets all necessary options to ensure
that the resulting message will be usable by a user of PGP 7.x.
* New --attribute-fd command for frontends and scripts to get the
contents of attribute packets (i.e. photos)
* In expert mode, the user can now re-sign a v3 key with a v4
self-signature. This does not change the v3 key into a v4 key,
but it does allow the user to use preferences, primary ID flags,
etc.
* Significantly improved photo ID support on non-unixlike
platforms.
* The version number has jumped ahead to 1.1.90 to skip over the
old version 1.1 and to get ready for the upcoming 1.2.
* ElGamal sign and encrypt is not anymore allowed in the key
generation dialog unless in expert mode. RSA sign and encrypt
has been added with the same restrictions.
* [W32] Keyserver access does work with Windows NT.
Noteworthy changes in version 1.0.7 (2002-04-29)
------------------------------------------------
* Secret keys are now stored and exported in a new format which
uses SHA-1 for integrity checks. This format renders the
Rosa/Klima attack useless. Other OpenPGP implementations might
not yet support this, so the option --simple-sk-checksum creates
the old vulnerable format.
* The default cipher algorithm for encryption is now CAST5,
default hash algorithm is SHA-1. This will give us better
interoperability with other OpenPGP implementations.
* Symmetric encrypted messages now use a fixed file size if
possible. This is a tradeoff: it breaks PGP 5, but fixes PGP 2,
6, and 7. Note this was only an issue with RFC-1991 style
symmetric messages.
* Photographic user ID support. This uses an external program to
view the images.
* Enhanced keyserver support via keyserver "plugins". GnuPG comes
with plugins for the NAI LDAP keyserver as well as the HKP email
keyserver. It retains internal support for the HKP HTTP
keyserver.
* Nonrevocable signatures are now supported. If a user signs a
key nonrevocably, this signature cannot be taken back so be
careful!
* Multiple signature classes are usable when signing a key to
specify how carefully the key information (fingerprint, photo
ID, etc) was checked.
* --pgp2 mode automatically sets all necessary options to ensure
that the resulting message will be usable by a user of PGP 2.x.
* --pgp6 mode automatically sets all necessary options to ensure
that the resulting message will be usable by a user of PGP 6.x.
* Signatures may now be given an expiration date. When signing a
key with an expiration date, the user is prompted whether they
want their signature to expire at the same time.
* Revocation keys (designated revokers) are now supported if
present. There is currently no way to designate new keys as
designated revokers.
* Permissions on the .gnupg directory and its files are checked
for safety.
* --expert mode enables certain silly things such as signing a
revoked user id, expired key, or revoked key.
* Some fixes to build cleanly under Cygwin32.
* New tool gpgsplit to split OpenPGP data formats into packets.
* New option --preserve-permissions.
* Subkeys created in the future are not used for encryption or
signing unless the new option --ignore-valid-from is used.
* Revoked user-IDs are not listed unless signatures are listed too
or we are in verbose mode.
* There is no default comment string with ascii armors anymore
except for revocation certificates and --enarmor mode.
* The command "primary" in the edit menu can be used to change the
primary UID, "setpref" and "updpref" can be used to change the
preferences.
* Fixed the preference handling; since 1.0.5 they were erroneously
matched against against the latest user ID and not the given one.
* RSA key generation.
* Merged Stefan's patches for RISC OS in. See comments in
scripts/build-riscos.
* It is now possible to sign and conventional encrypt a message (-cs).
* The MDC feature flag is supported and can be set by using
the "updpref" edit command.
* The status messages GOODSIG and BADSIG are now returning the primary
UID, encoded using %XX escaping (but with spaces left as spaces,
so that it should not break too much)
* Support for GDBM based keyrings has been removed.
* The entire keyring management has been revamped.
* The way signature stati are store has changed so that v3
signatures can be supported. To increase the speed of many
operations for existing keyrings you can use the new
--rebuild-keydb-caches command.
* The entire key validation process (trustdb) has been revamped.
See the man page entries for --update-trustdb, --check-trustdb
and --no-auto-check-trustdb.
* --trusted-keys is again obsolete, --edit can be used to set the
ownertrust of any key to ultimately trusted.
* A subkey is never used to sign keys.
* Read only keyrings are now handled as expected.
Noteworthy changes in version 1.0.6 (2001-05-29)
------------------------------------------------
* Security fix for a format string bug in the tty code.
* Fixed format string bugs in all PO files.
* Removed Russian translation due to too many bugs. The FTP
server has an unofficial but better translation in the contrib
directory.
* Fixed expire time calculation and keyserver access.
* The usual set of minor bug fixes and enhancements.
* non-writable keyrings are now correctly handled.
Noteworthy changes in version 1.0.5 (2001-04-29)
------------------------------------------------
* WARNING: The semantics of --verify have changed to address a
problem with detached signature detection. --verify now ignores
signed material given on stdin unless this is requested by using
a "-" as the name for the file with the signed material. Please
check all your detached signature handling applications and make
sure that they don't pipe the signed material to stdin without
using a filename together with "-" on the the command line.
* WARNING: Corrected hash calculation for input data larger than
512M - it was just wrong, so you might notice bad signature in
some very big files. It may be wise to keep an old copy of
GnuPG around.
* Secret keys are no longer imported unless you use the new option
--allow-secret-key-import. This is a kludge and future versions will
handle it in another way.
* New command "showpref" in the --edit-key menu to show an easier
to understand preference listing.
* There is now the notation of a primary user ID. For example, it
is printed with a signature verification as the first user ID;
revoked user IDs are not printed there anymore. In general the
primary user ID is the one with the latest self-signature.
* New --charset=utf-8 to bypass all internal conversions.
* Large File Support (LFS) is now working.
* New options: --ignore-crc-error, --no-sig-create-check,
--no-sig-cache, --fixed_list_mode, --no-expensive-trust-checks,
--enable-special-filenames and --use-agent. See man page.
* New command --pipemode, which can be used to run gpg as a
co-process. Currently only the verification of detached
signatures are working. See doc/DETAILS.
* Keyserver support for the W32 version.
* Rewritten key selection code so that GnuPG can better cope with
multiple subkeys, expire dates and so. The drawback is that it
is slower.
* A whole lot of bug fixes.
* The verification status of self-signatures are now cached. To
increase the speed of key list operations for existing keys you
can do the following in your GnuPG homedir (~/.gnupg):
cp pubring.gpg pubring.gpg.save && gpg --export-all >x && \
rm pubring.gpg && gpg --import x
Only v4 keys (i.e not the old RSA keys) benefit from this caching.
* New translations: Estonian, Turkish.
Noteworthy changes in version 1.0.4 (2000-10-17)
------------------------------------------------
* Fixed a serious bug which could lead to false signature verification
results when more than one signature is fed to gpg. This is the
primary reason for releasing this version.
* New utility gpgv which is a stripped down version of gpg to
be used to verify signatures against a list of trusted keys.
* Rijndael (AES) is now supported and listed with top preference.
* --with-colons now works with --print-md[s].
Noteworthy changes in version 1.0.3 (2000-09-18)
------------------------------------------------
* Fixed problems with piping to/from other MS-Windows software
@ -17,6 +260,10 @@ Noteworthy changes in version 1.1.2
* Revoked user IDs are now marked in the output of --list-key
* New options --show-session-key and --override-session-key
to help the British folks to somewhat minimize the danger
of this Orwellian RIP bill.
* New options --merge-only and --try-all-secrets.
* New configuration option --with-egd-socket.
@ -24,7 +271,7 @@ Noteworthy changes in version 1.1.2
* The --trusted-key option is back after it left us with 0.9.5
* RSA is supported. Key generation does not yet work but will come
soon.
soon.
* CAST5 and SHA-1 are now the default algorithms to protect the key
and for symmetric-only encryption. This should solve a couple
@ -34,14 +281,10 @@ Noteworthy changes in version 1.1.2
* Twofish and MDC enhanced encryption is now used. PGP 7 supports
this. Older versions of GnuPG don't support it, so they should be
upgraded to at least 1.0.2
Noteworthy changes in version 1.1.1
-----------------------------------
* Add gpg-agent.
* Removed option --emulate-checksum-bug
Noteworthy changes in version 1.0.2 (2000-07-12)
----------------------------------------------
* Fixed expiration handling of encryption keys.
@ -93,7 +336,12 @@ Noteworthy changes in version 1.1.1
entirely. This option should not yet be used.
* New option --no-auto-key-retrieve to disable retrieving of
a missing public key from a keyerver, when a keyerver has been set.
a missing public key from a keyserver, when a keyserver has been set.
* Danish translation
Noteworthy changes in version 1.0.1 (1999-12-16)
-----------------------------------
* New command --verify-files. New option --fast-list-mode.
@ -102,7 +350,7 @@ Noteworthy changes in version 1.1.1
* Fixed some minor bugs and the problem with conventional encrypted
packets which did use the gpg v3 partial length headers.
* Some more translations.
* Add Indonesian and Portugese translations.
* Fixed a bug with symmetric-only encryption using the non-default 3DES.
The option --emulate-3des-s2k-bug may be used to decrypt documents
@ -120,16 +368,8 @@ Noteworthy changes in version 1.1.1
* New keys are now generated with an additional preference to Blowfish.
* Removed the GNU Privacy Handbook from the distribution because it
is now in the package GPH.
Noteworthy changes in version 1.1.0 (1999-10-26)
-----------------------------------
* Did a couple of changes for this new development series.
This release basically works on my machine but may have
serious problems.
* Removed the GNU Privacy Handbook from the distribution as it will go
into a separate one.
Noteworthy changes in version 1.0.0 (1999-09-07)
@ -928,3 +1168,13 @@ Noteworthy changes in version 0.2.3
* The string "(INSECURE!)" is appended to a new user-id if this
is generated on a system without a good random number generator.
Copyright 1998, 1999, 2000, 2001, 2002 Free Software Foundation, Inc.
This file is free software; as a special exception the author gives
unlimited permission to copy and/or distribute it, with or without
modifications, as long as this notice is preserved.
This file is distributed in the hope that it will be useful, but
WITHOUT ANY WARRANTY, to the extent permitted by law; without even the
implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.

36
NOTES
View File

@ -1,7 +1,6 @@
Checking ElGamal signatures is really slow and the reason for the long
running time on parts o my keyring. Because somekeys are also checked at startup, this is even worser. I should invalidate my self signature with algo 16 or 20.
gcry_md_open() may return NULL in case of an unknown algorithm
Return GCRYERR_NOMEM instead of using m_alloc in most library functions
SCO UnixWare/7.1.0 reported by Allan Clark <allanc@sco.com> for 0.9.8
@ -13,16 +12,35 @@ Some other reported cpu-vendor-os strings:
sparc-sun-solaris5.4
sparc-sun-sunos4.1.2
i386-pc-sysv4.2 (USL Unixware v1.1.2)
powerpc-ibm-aix4.3.2.0 John Payne <jcapayne@att.com>
gpg 1.0.1 okay with MP-RAS 3.02.01 Edition 5 using gcc 2.95.2 and EGD
By <CSpeicher@eisi.com>
gpg 1.0.1 okay with 4.0.1 BSDI BSD/OS 4.0 i386
rndw32 tested on:
Windows 98 4.10.1998 mit einem AMD-K6-2-450
Michael Engels <angel@dalrin.de>)
Windows 95 4.00.950a
Windows NT 4.00.1381
tried to compile GnuPG on AIX 4.3 on a power CPU based machine. It
doesn't work out of the box but i found a way to do so:
Simos Hadjiyiannis <hadjiyia@cs.colostate.edu>
runs successfully on HP-UX v11.00 as well. (using gcc)
Johan Wevers <johanw@vulcan.xs4all.nl> is working on a NL translation.
PowerPC based machines:
CFLAGS="-g -O2 -mcpu=powerpc" ./configure --disable-asm --disable-dynload
+--enable-static-rnd=egd
Power1 and Power2 machines:
CFLAGS="-g -O2 -mcpu=power" ./configure --disable-asm --disable-dynload
+--enable-static-rnd=egd

32
OBUGS
View File

@ -100,35 +100,3 @@
Ctrl-D does not work correct for messages entered at the tty.
FIX: 1999-06-18 (Better EOF detection on terminals)
[ *] #22 1999-07-22
Solaris make has problems with the generated POTFILES - seems to be a
gettext bug. Use GNU make as a workaround.
FIX: 1999-12-03 (meanwhile fixed in gettext)
[ *] #23 1999-09-03 <todd.brooks@yale.edu> 0.9.11
Only the first signature of a cleartext sig seems to be verified.
Can't fix it in 1.0 because the code is semi-frozen.
HMMM: Can't reprodude the bug here - it just works.
[ **] #25 1999-10-04 <wk@gnupg.org> 1.0
Validity check based on user name is only done when listing a key;
the key ID based checking (selecting the user ID with the most
validity) is used in all other cases. The Edit menu does not have
a way to disable user ID based validity.
FIX: 1999-12-03 (1.0.x only)
[ *] #26 1999-11-11
gpg still does not take UTF8 strings to select a UID.
FIX: 1999-11-12
[ **] #27 1999-11-12 <oren@hishome.net>
Unknown packets (type 17 - photo ID?) mess up the checking
of self-signature because they are simply ignored.
FIX: 1999-11-12 (1.0.x only)
[ **] #29 1999-12-01 <karney@pppl.gov>
Using a --gen-key on a new user account (w/o ~./gnupg) doesn't pay
attention to the standard options files, which get's installed
afterwards.
FIX: 1999-12-02

20
PROJECTS
View File

@ -15,8 +15,9 @@
* add an option to re-create a public key from a secret key; we
can do this in trustdb.c:verify_own_keys. (special tool?)
* rewrite --list-packets or put it into another tool.
Hmmm, we better drop the duplication of the public part and just keep
the secrets in the "secring" - this has the additional that we can
put those secrets on a hardware token.
* write a tool to extract selected keys from a file.
@ -28,3 +29,18 @@
We should add an abstraction layer so that adding support for
different certificate structures will become easier.
* "Michael T. Babcock" <mbabcock@fibrespeed.net> suggested to write
an event log so that other software can display a key history or
alike with GnuPG results. This should be connected to the keyrings.
Copyright 1998, 1999, 2000, 2001 Free Software Foundation, Inc.
This file is free software; as a special exception the author gives
unlimited permission to copy and/or distribute it, with or without
modifications, as long as this notice is preserved.
This file is distributed in the hope that it will be useful, but
WITHOUT ANY WARRANTY, to the extent permitted by law; without even the
implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.

500
README
View File

@ -1,17 +1,497 @@
GnuPG - The GNU Privacy Guard
-------------------------------
Version 1.1
Version 1.0
WARNING: This is the current development branch
of GnuPG. THIS SHOULD NOT BE USED IN
A PRODUCTION ENVIRONMENT. It will
change quite often and may have serious
problems. Use the GnuPG from the stable
Branch 1.0.x for real work. The next
stable release will be 1.2
Copyright 1998, 1999, 2000, 2001, 2002 Free Software Foundation, Inc.
If you are looking for the CVS from the stable tree, use
This file is free software; as a special exception the author gives
unlimited permission to copy and/or distribute it, with or without
modifications, as long as this notice is preserved.
cvs -z3 -d :pserver:anoncvs@cvs.gnupg.org:/cvs/gnupg co -r STABLE-BRANCH-1-0 gnupg
This file is distributed in the hope that it will be useful, but
WITHOUT ANY WARRANTY, to the extent permitted by law; without even the
implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
Intro
-----
GnuPG is GNU's tool for secure communication and data storage.
It can be used to encrypt data and to create digital signatures.
It includes an advanced key management facility and is compliant
with the proposed OpenPGP Internet standard as described in RFC2440.
GnuPG works best on GNU/Linux or *BSD systems. Most other Unices
are also supported but are not as well tested as the Free Unices.
See http://www.gnupg.org/gnupg.html#supsys for a list of systems
which are known to work.
See the file COPYING for copyright and warranty information.
Because GnuPG does not use use any patented algorithm it cannot be
compatible with PGP2 versions. PGP 2.x uses IDEA (which is patented
worldwide).
The default algorithms are DSA and ElGamal. ElGamal for signing
is still available, but because of the larger size of such
signatures it is deprecated (Please note that the GnuPG
implementation of ElGamal signatures is *not* insecure). Symmetric
algorithms are: AES, 3DES, Blowfish, CAST5 and Twofish
Digest algorithms available are MD5, RIPEMD160 and SHA1.
Installation
------------
Please read the file INSTALL!
Here is a quick summary:
1) Check that you have unmodified sources. The below on how to do this.
Don't skip it - this is an important step!
2) Unpack the TAR. With GNU tar you can do it this way:
"tar xzvf gnupg-x.y.z.tar.gz"
3) "cd gnupg-x.y.z"
4) "./configure"
5) "make"
6) "make install"
7) You end up with a "gpg" binary in /usr/local/bin.
Note: Because some old programs rely on the existence of a
binary named "gpgm" (which was build by some Beta versions
of GnuPG); you may want to install a symbolic link to it:
"cd /usr/local/bin; ln -s gpg gpgm"
8) To avoid swapping out of sensitive data, you can install "gpg" as
suid root. If you don't do so, you may want to add the option
"no-secmem-warning" to ~/.gnupg/options
How to Verify the Source
------------------------
In order to check that the version of GnuPG which you are going to
install is an original and unmodified one, you can do it in one of
the following ways:
a) If you already have a trusted Version of GnuPG installed, you
can simply check the supplied signature:
$ gpg --verify gnupg-x.y.z.tar.gz.asc
This checks that the detached signature gnupg-x.y.z.tar.gz.asc
is indeed a a signature of gnupg-x.y.z.tar.gz. The key used to
create this signature is:
"pub 1024D/57548DCD 1998-07-07 Werner Koch (gnupg sig) <dd9jn@gnu.org>"
If you do not have this key, you can get it from the source in
the file doc/samplekeys.asc (use "gpg --import doc/samplekeys.asc"
to add it to the keyring) or from any keyserver. You have to
make sure that this is really the key and not a faked one. You
can do this by comparing the output of:
$ gpg --fingerprint 0x57548DCD
with the elsewhere published fingerprint
Please note, that you have to use an old version of GnuPG to
do all this stuff. *Never* use the version which you are going
to check!
b) If you don't have any of the above programs, you have to verify
the MD5 checksum:
$ md5sum gnupg-x.y.z.tar.gz
This should yield an output _similar_ to this:
fd9351b26b3189c1d577f0970f9dcadc gnupg-x.y.z.tar.gz
Now check that this checksum is _exactly_ the same as the one
published via the announcement list and probably via Usenet.
Documentation
-------------
The manual will be distributed separate under the name "gph".
An online version of the latest manual draft is available at the
GnuPG web pages:
http://www.gnupg.org/gph/
A list of frequently asked questions is available in GnuPG's
distibution in the file doc/FAQ and online as:
http://www.gnupg.org/faq.html
A couple of HOWTO documents are available online; for a listing see:
http://www.gnupg.org/docs.html#howtos
A man page with a description of all commands and options gets installed
along with the program.
Introduction
------------
Here is a brief overview on how to use GnuPG - it is strongly suggested
that you read the manual and other information about the use of
cryptography. GnuPG is only a tool, secure usage requires that
YOU KNOW WHAT YOU ARE DOING.
If you already have a DSA key from PGP 5 (they call them DH/ElGamal)
you can simply copy the pgp keyrings over the GnuPG keyrings after
running gpg once to create the correct directory.
The normal way to create a key is
gpg --gen-key
This asks some questions and then starts key generation. To create
good random numbers for the key parameters, GnuPG needs to gather
enough noise (entropy) from your system. If you see no progress
during key generation you should start some other activities such
as mouse moves or hitting on the CTRL and SHIFT keys.
Generate a key ONLY on a machine where you have direct physical
access - don't do it over the network or on a machine used also
by others - especially if you have no access to the root account.
When you are asked for a passphrase use a good one which you can
easy remember. Don't make the passphrase too long because you have
to type it for every decryption or signing; but, - AND THIS IS VERY
IMPORTANT - use a good one that is not easily to guess because the
security of the whole system relies on your secret key and the
passphrase that protects it when someone gains access to your secret
keyring. A good way to select a passphrase is to figure out a short
nonsense sentence which makes some sense for you and modify it by
inserting extra spaces, non-letters and changing the case of some
characters - this is really easy to remember especially if you
associate some pictures with it.
Next, you should create a revocation certificate in case someone
gets knowledge of your secret key or you forgot your passphrase
gpg --gen-revoke your_user_id
Run this command and store the revocation certificate away. The output
is always ASCII armored, so that you can print it and (hopefully
never) re-create it if your electronic media fails.
Now you can use your key to create digital signatures
gpg -s file
This creates a file "file.gpg" which is compressed and has a
signature attached.
gpg -sa file
Same as above, but creates a file "file.asc" which is ASCII armored
and and ready for sending by mail. It is better to use your
mailers features to create signatures (The mailer uses GnuPG to do
this) because the mailer has the ability to MIME encode such
signatures - but this is not a security issue.
gpg -s -o out file
Creates a signature of "file", but writes the output to the file
"out".
Everyone who knows your public key (you can and should publish
your key by putting it on a key server, a web page or in your .plan
file) is now able to check whether you really signed this text
gpg --verify file
GnuPG now checks whether the signature is valid and prints an
appropriate message. If the signature is good, you know at least
that the person (or machine) has access to the secret key which
corresponds to the published public key.
If you run gpg without an option it will verify the signature and
create a new file that is identical to the original. gpg can also
run as a filter, so that you can pipe data to verify trough it
cat signed-file | gpg | wc -l
which will check the signature of signed-file and then display the
number of lines in the original file.
To send a message encrypted to someone you can use
gpg -e -r heine file
This encrypts "file" with the public key of the user "heine" and
writes it to "file.gpg"
echo "hello" | gpg -ea -r heine | mail heine
Ditto, but encrypts "hello\n" and mails it as ASCII armored message
to the user with the mail address heine.
gpg -se -r heine file
This encrypts "file" with the public key of "heine" and writes it
to "file.gpg" after signing it with your user id.
gpg -se -r heine -u Suttner file
Ditto, but sign the file with your alternative user id "Suttner"
GnuPG has some options to help you publish public keys. This is
called "exporting" a key, thus
gpg --export >all-my-keys
exports all the keys in the keyring and writes them (in a binary
format) to "all-my-keys". You may then mail "all-my-keys" as an
MIME attachment to someone else or put it on an FTP server. To
export only some user IDs, you give them as arguments on the command
line.
To mail a public key or put it on a web page you have to create
the key in ASCII armored format
gpg --export --armor | mail panther@tiger.int
This will send all your public keys to your friend panther.
If you have received a key from someone else you can put it
into your public keyring. This is called "importing"
gpg --import [filenames]
New keys are appended to your keyring and already existing
keys are updated. Note that GnuPG does not import keys that
are not self-signed.
Because anyone can claim that a public key belongs to her
we must have some way to check that a public key really belongs
to the owner. This can be achieved by comparing the key during
a phone call. Sure, it is not very easy to compare a binary file
by reading the complete hex dump of the file - GnuPG (and nearly
every other program used for management of cryptographic keys)
provides other solutions.
gpg --fingerprint <username>
prints the so called "fingerprint" of the given username which
is a sequence of hex bytes (which you may have noticed in mail
sigs or on business cards) that uniquely identifies the public
key - different keys will always have different fingerprints.
It is easy to compare fingerprints by phone and I suggest
that you print your fingerprint on the back of your business
card. To see the fingerprints of the secondary keys, you can
give the command twice; but this is normally not needed.
If you don't know the owner of the public key you are in trouble.
Suppose however that friend of yours knows someone who knows someone
who has met the owner of the public key at some computer conference.
Suppose that all the people between you and the public key holder
may now act as introducers to you. Introducers signing keys thereby
certify that they know the owner of the keys they sign. If you then
trust all the introducers to have correctly signed other keys, you
can be be sure that the other key really belongs to the one who
claims to own it..
There are 2 steps to validate a key:
1. First check that there is a complete chain
of signed keys from the public key you want to use
and your key and verify each signature.
2. Make sure that you have full trust in the certificates
of all the introduces between the public key holder and
you.
Step 2 is the more complicated part because there is no easy way
for a computer to decide who is trustworthy and who is not. GnuPG
leaves this decision to you and will ask you for a trust value
(here also referenced as the owner-trust of a key) for every key
needed to check the chain of certificates. You may choose from:
a) "I don't know" - then it is not possible to use any
of the chains of certificates, in which this key is used
as an introducer, to validate the target key. Use this if
you don't know the introducer.
b) "I do not trust" - Use this if you know that the introducer
does not do a good job in certifying other keys. The effect
is the same as with a) but for a) you may later want to
change the value because you got new information about this
introducer.
c) "I trust marginally" - Use this if you assume that the
introducer knows what he is doing. Together with some
other marginally trusted keys, GnuPG validates the target
key then as good.
d) "I fully trust" - Use this if you really know that this
introducer does a good job when certifying other keys.
If all the introducer are of this trust value, GnuPG
normally needs only one chain of signatures to validate
a target key okay. (But this may be adjusted with the help
of some options).
This information is confidential because it gives your personal
opinion on the trustworthiness of someone else. Therefore this data
is not stored in the keyring but in the "trustdb"
(~/.gnupg/trustdb.gpg). Do not assign a high trust value just
because the introducer is a friend of yours - decide how well she
understands the implications of key signatures and you may want to
tell her more about public key cryptography so you can later change
the trust value you assigned.
Okay, here is how GnuPG helps you with key management. Most stuff
is done with the --edit-key command
gpg --edit-key <keyid or username>
GnuPG displays some information about the key and then prompts
for a command (enter "help" to see a list of commands and see
the man page for a more detailed explanation). To sign a key
you select the user ID you want to sign by entering the number
that is displayed in the leftmost column (or do nothing if the
key has only one user ID) and then enter the command "sign" and
follow all the prompts. When you are ready, give the command
"save" (or use "quit" to cancel your actions).
If you want to sign the key with another of your user IDs, you
must give an "-u" option on the command line together with the
"--edit-key".
Normally you want to sign only one user ID because GnuPG
uses only one and this keeps the public key certificate
small. Because such key signatures are very important you
should make sure that the signatories of your key sign a user ID
which is very likely to stay for a long time - choose one with an
email address you have full control of or do not enter an email
address at all. In future GnuPG will have a way to tell which
user ID is the one with an email address you prefer - because
you have no signatures on this email address it is easy to change
this address. Remember, your signatories sign your public key (the
primary one) together with one of your user IDs - so it is not possible
to change the user ID later without voiding all the signatures.
Tip: If you hear about a key signing party on a computer conference
join it because this is a very convenient way to get your key
certified (But remember that signatures have nothing to to with the
trust you assign to a key).
8 Ways to Specify a User ID
--------------------------
There are several ways to specify a user ID, here are some examples.
* Only by the short keyid (prepend a zero if it begins with A..F):
"234567C4"
"0F34E556E"
"01347A56A"
"0xAB123456
* By a complete keyid:
"234AABBCC34567C4"
"0F323456784E56EAB"
"01AB3FED1347A5612"
"0x234AABBCC34567C4"
* By a fingerprint:
"1234343434343434C434343434343434"
"123434343434343C3434343434343734349A3434"
"0E12343434343434343434EAB3484343434343434"
The first one is MD5 the others are ripemd160 or sha1.
* By an exact string:
"=Heinrich Heine <heinrichh@uni-duesseldorf.de>"
* By an email address:
"<heinrichh@uni-duesseldorf.de>"
* By word match
"+Heinrich Heine duesseldorf"
All words must match exactly (not case sensitive) and appear in
any order in the user ID. Words are any sequences of letters,
digits, the underscore and characters with bit 7 set.
* Or by the usual substring:
"Heine"
"*Heine"
The '*' indicates substring search explicitly.
Batch mode
----------
If you use the option "--batch", GnuPG runs in non-interactive mode and
never prompts for input data. This does not even allow entering the
passphrase. Until we have a better solution (something like ssh-agent),
you can use the option "--passphrase-fd n", which works like PGP's
PGPPASSFD.
Batch mode also causes GnuPG to terminate as soon as a BAD signature is
detected.
Exit status
-----------
GnuPG returns with an exit status of 1 if in batch mode and a bad signature
has been detected or 2 or higher for all other errors. You should parse
stderr or, better, the output of the fd specified with --status-fd to get
detailed information about the errors.
How to Get More Information
---------------------------
The primary WWW page is "http://www.gnupg.org"
The primary FTP site is "ftp://ftp.gnupg.org/gcrypt/"
See http://www.gnupg.org/mirrors.html for a list of mirrors
and use them if possible. You may also find GnuPG mirrored on
some of the regular GNU mirrors.
We have some mailing lists dedicated to GnuPG:
gnupg-announce@gnupg.org For important announcements like
new versions and such stuff.
This is a moderated list and has
very low traffic.
gnupg-users@gnupg.org For general user discussion and
help.
gnupg-devel@gnupg.org GnuPG developers main forum.
You subscribe to one of the list by sending mail with a subject
of "subscribe" to x-request@gnupg.org, where x is the name of the
mailing list (gnupg-announce, gnupg-users, etc.). An archive of
the mailing lists is available at http://lists.gnupg.org .
The gnupg.org domain is hosted in Germany to avoid possible legal
problems (technical advices may count as a violation of ITAR).
Please direct bug reports to <gnupg-bugs@gnu.org> or post
them direct to the mailing list <gnupg-devel@gnupg.org>.
Please direct questions about GnuPG to the users mailing list or
one of the pgp newsgroups; please do not direct questions to one
of the authors directly as we are busy working on improvements
and bug fixes. Both mailing lists are watched by the authors
and we try to answer questions when time allows us to do so.
Commercial grade support for GnuPG is available; please see
the GNU service directory or search other resources.

123
THANKS
View File

@ -3,79 +3,154 @@ reporting problems, suggesting various improvements or submitting actual
code. Here is a list of those people. Help me keep it complete and free of
errors.
Adam Mitchell adam@cafe21.org
Albert Chin china@thewrittenword.com
Alec Habig habig@budoe2.bu.edu
Allan Clark allanc@sco.com
Anand Kumria wildfire@progsoc.uts.edu.au
Andreas Haumer andreas@xss.co.at
Anthony Mulcahy anthony@kcn.ne.jp
Ariel T Glenn ariel@columbia.edu
Bodo Moeller Bodo_Moeller@public.uni-hamburg.de
Brendan O'Dea bod@debian.org
Brenno de Winter brenno@dewinter.com
Brian M. Carlson karlsson@hal-pc.org
Brian Moore bem@cmc.net
Brian Warner warner@lothar.com
Bryan Fullerton bryanf@samurai.com
Caskey L. Dickson caskey@technocage.com
Cees van de Griend cees-list@griend.xs4all.nl
Charles Levert charles@comm.polymtl.ca
Chip Salzenberg chip@valinux.com
Chris Adams cmadams@hiwaay.net
Christian Kurz shorty@debian.org
Christian von Roques roques@pond.sub.org
Christopher Oliver oliver@fritz.traverse.net
Christian Recktenwald chris@citecs.de
Dan Winship danw@helixcode.com
Daniel Eisenbud eisenbud@cs.swarthmore.edu
Daniel Koening dan@mail.isis.de
Daniel Resare daniel@resare.com
Dave Dykstra dwd@bell-labs.com
David C Niemi niemi@tuxers.net
David Champion dgc@uchicago.edu
David Ellement ellement@sdd.hp.com
David Hallinan hallinan@rtd.com
David Hollenberg dhollen@ISI.EDU
David Mathog MATHOG@seqaxp.bio.caltech.edu
David R. Bergstein dbergstein@home.com
David Shaw dshaw@jabberwocky.com
Detlef Lannert lannert@lannert.rz.uni-duesseldorf.de
Dimitri dmitri@advantrix.com
Dirk Lattermann dlatt@t-online.de
Dirk Meyer dirk.meyer@dinoex.sub.org
Disastry Disastry@saiknes.lv
Douglas Calvert dfc@anize.org
Ed Boraas ecxjo@esperanto.org
Edmund GRIMLEY EVANS edmundo@rano.org
Edwin Woudt edwin@woudt.nl
Enzo Michelangeli em@MailAndNews.com
Ernst Molitor ernst.molitor@uni-bonn.de
Fabio Coatti cova@felix.unife.it
Fabio Coatti cova@ferrara.linux.it
Felix von Leitner leitner@amdiv.de
fish stiqz fish@analog.org
Florian Weimer Florian.Weimer@rus.uni-stuttgart.de
Francesco Potorti pot@gnu.org
Frank Donahoe fdonahoe@wilkes1.wilkes.edu
Frank Heckenbach heckenb@mi.uni-erlangen.de
Frank Stajano frank.stajano@cl.cam.ac.uk
Gaël Quéri gqueri@mail.dotcom.fr
Frank Tobin ftobin@uiuc.edu
Gabriel Rosenkoetter gr@eclipsed.net
Gaël Quéri gael@lautre.net
Gene Carter gcarter@lanier.com
Georg Schwarz georg.schwarz@iname.com
Giampaolo Tomassoni g.tomassoni@libero.it
Gilbert Fernandes gilbert_fernandes@hotmail.com
Greg Louis glouis@dynamicro.on.ca
Greg Troxel gdt@ir.bbn.com
Gregory Steuck steuck@iname.com
Geoff Keating geoffk@ozemail.com.au
Harald Denker harry@hal.westfalen.de
Holger Baust Holger.Baust@freenet-ag.de
Hendrik Buschkamp buschkamp@rheumanet.org
Holger Schurig holger@d.om.org
Holger Smolinski smolinsk@de.ibm.com
Holger Trapp Holger.Trapp@informatik.tu-chemnitz.de
Hugh Daniel hugh@toad.com
Huy Le huyle@ugcs.caltech.edu
Ian McKellar imckellar@harvestroad.com.au
Ivo Timmermans itimmermans@bigfoot.com
Jan Krueger max@physics.otago.ac.nz
Jan Niehusmann jan@gondor.com
Janusz A. Urbanowicz alex@bofh.torun.pl
James Troup james@nocrew.org
Jean-loup Gailly gzip@prep.ai.mit.edu
Jeff Long long@kestrel.cc.ukans.edu
Jeffery Von Ronne jronne@ics.uci.edu
Jens Bachem bachem@rrz.uni-koeln.de
Jeroen C. van Gelderen jeroen@vangelderen.org
J Horacio MG homega@ciberia.es
J. Michael Ashley jashley@acm.org
Jim Bauer jfbauer@home.com
Jim Small cavenewt@my-deja.com
Joachim Backes backes@rhrk.uni-kl.de
Joe Rhett jrhett@isite.net
John A. Martin jam@jamux.com
Johnny Teveßen j.tevessen@gmx.de
Jörg Schilling schilling@fokus.gmd.de
Jos Backus Jos.Backus@nl.origin-it.com
Jun Kuriyama kuriyama@sky.rim.or.jp
Kahil D. Jallad kdj4@cs.columbia.edu
Karl Fogel kfogel@guanabana.onshore.com
Karsten Thygesen karthy@kom.auc.dk
Katsuhiro Kondou kondou@nec.co.jp
Kazu Yamamoto kazu@iijlab.net
Keith Clayton keith@claytons.org
Kevin Ryde user42@zip.com.au
Klaus Singvogel ks@caldera.de
Kurt Garloff garloff@suse.de
Lars Kellogg-Stedman lars@bu.edu
L. Sassaman rabbi@quickie.net
M Taylor mctaylor@privacy.nb.ca
Marcel Waldvogel mwa@arl.wustl.edu
Marco d'Itri md@linux.it
Marcus Brinkmann Marcus.Brinkmann@ruhr-uni-bochum.de
Mark Adler madler@alumni.caltech.edu
Mark Elbrecht snowball3@bigfoot.com
Mark Pettit pettit@yahoo-inc.com
Markus Friedl Markus.Friedl@informatik.uni-erlangen.de
Martin Kahlert martin.kahlert@provi.de
Martin Hamilton
Martin Schulte schulte@thp.uni-koeln.de
Matt Kraai kraai@alumni.carnegiemellon.edu
Matthew Skala mskala@ansuz.sooke.bc.ca
Matthew Wilcox matthew@wil.cx
Matthias Urlichs smurf@noris.de
Max Valianskiy maxcom@maxcom.ml.org
Michael Engels michael.engels@uni-duesseldorf.de
Michael Fischer v. Mollard mfvm@gmx.de
Michael Roth mroth@nessie.de
Michael Sobolev mss@despair.transas.com
Michael Tokarev mjt@tls.msk.ru
Nicolas Graner Nicolas.Graner@cri.u-psud.fr
Mike McEwan mike@lotusland.demon.co.uk
Neal H Walfield neal@cs.uml.edu
NIIBE Yutaka gniibe@chroot.org
Niklas Hernaeus
Nimrod Zimerman zimerman@forfree.at
N J Doye nic@niss.ac.uk
Oliver Haakert haakert@hsp.de
Oskari Jääskeläinen f33003a@cc.hut.fi
Pascal Scheffers Pascal@scheffers.net
Paul D. Smith psmith@baynetworks.com
Per Cederqvist ceder@lysator.liu.se
Phil Blundell pb@debian.org
Philippe Laliberte arsphl@oeil.qc.ca
Peter Fales psfales@lucent.com
Peter Gutmann pgut001@cs.auckland.ac.nz
Peter Marschall Peter.Marschall@gedos.de
Peter Valchev pvalchev@openbsd.org
Piotr Krukowiecki piotr@pingu.ii.uj.edu.pl
QingLong qinglong@bolizm.ihep.su
Ralph Gillen gillen@theochem.uni-duesseldorf.de
Rat ratinox@peorth.gweep.net
@ -83,33 +158,69 @@ Reinhard Wobst R.Wobst@ifw-dresden.de
Rémi Guyomarch rguyom@mail.dotcom.fr
Reuben Sumner rasumner@wisdom.weizmann.ac.il
Richard Outerbridge outer@interlog.com
Robert Joop rj@rainbow.in-berlin.de
Roddy Strachan roddy@satlink.com.au
Roger Sondermann r.so@bigfoot.com
Roland Rosenfeld roland@spinnaker.rhein.de
Roman Pavlik rp@tns.cz
Ross Golder rossigee@bigfoot.com
Sam Roberts sam@cogent.ca
Sami Tolvanen sami@tolvanen.com
Sean MacLennan seanm@netwinder.org
Sebastian Klemke packet@convergence.de
Serge Munhoven munhoven@mema.ucl.ac.be
SL Baur steve@xemacs.org
Stefan Bellon sbellon@sbellon.de
Stefan Karrmann S.Karrmann@gmx.net
Stefan Keller dres@cs.tu-berlin.de
Steffen Ullrich ccrlphr@xensei.com
Steffen Zahn zahn@berlin.snafu.de
Steven Bakker steven@icoe.att.com
Steven Murdoch sjmurdoch@bigfoot.com
Susanne Schultz schultz@hsp.de
Ted Cabeen secabeen@pobox.com
Thiago Jung Bauermann jungmann@cwb.matrix.com.br
Thomas Roessler roessler@guug.de
Tim Mooney mooney@dogbert.cc.ndsu.nodak.edu
Timo Schulz towaday@freakmail.de
Tom Spindler dogcow@home.merit.edu
Tom Zerucha tzeruch@ceddec.com
Tomas Fasth tomas.fasth@twinspot.net
Tommi Komulainen Tommi.Komulainen@iki.fi
Thomas Klausner wiz@danbala.ifoer.tuwien.ac.at
Tomasz Kozlowski tomek@rentec.com
Thomas Mikkelsen tbm@image.dk
Ulf Möller 3umoelle@informatik.uni-hamburg.de
Urko Lusa ulusa@euskalnet.net
Vincent P. Broman broman@spawar.navy.mil
W Lewis wiml@hhhh.org
Walter Hofmann Walter.Hofmann@physik.stud.uni-erlangen.de
Walter Koch koch@hsp.de
Werner Koch werner.koch@guug.de
Wayne Chapeskie waynec@spinnaker.com
Werner Koch wk@gnupg.org
Wim Vandeputte bunbun@reptile.rug.ac.be
Winona Brown win@huh.org
Yosiaki IIDA iida@ring.gr.jp
Yoshihiro Kajiki kajiki@ylug.org
disastry@saiknes.lv
nbecker@hns.com
Thanks to the German Unix User Group for providing FTP space,
Martin Hamilton for initially hosting the mailing list, HSP for
hosting gnupg.org and Linux Laptops Ltd. for a nice toy.
Thanks to the German Unix User Group for sponsoring this project,
Martin Hamilton for hosting the first mailing list and OpenIT for
cheap hosting conditions.
The development of this software has partly been funded by the German
Ministry for Economics and Technology under grant VIB3-68553.168-001/1999.
Many thanks to my wife Gerlinde for having so much patience with
me while hacking late in the evening.
Copyright 1998, 1999, 2000, 2001 Free Software Foundation, Inc.
This file is free software; as a special exception the author gives
unlimited permission to copy and/or distribute it, with or without
modifications, as long as this notice is preserved.
This file is distributed in the hope that it will be useful, but
WITHOUT ANY WARRANTY, to the extent permitted by law; without even the
implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.

9
THOUGHTS
View File

@ -264,3 +264,12 @@ Suggested things which I will not do:
--> Use a script and --status-fd
Copyright 1998, 1999, 2000, 2001 Free Software Foundation, Inc.
This file is free software; as a special exception the author gives
unlimited permission to copy and/or distribute it, with or without
modifications, as long as this notice is preserved.
This file is distributed in the hope that it will be useful, but
WITHOUT ANY WARRANTY, to the extent permitted by law; without even the
implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.

157
TODO
View File

@ -1,83 +1,124 @@
* Reword the "Not enough entropy" messages.
* Look at FEATURES packet to decide wether to use MDC. We should
rethink the whole preference implementation and drop that ugly hamehash
stuff. Is it possible to that in getkey? We have to see whether we
realy lookit up by user name or whethre there is a double lookup,
first by userid and then later by keyid - which is bad.
* Do we need a configure test for putenv?
* options.skel wird mit umask von root installiert.
* Check for consistent spelling of user ID, key ID etc.
Replace "user id not found" in getkey.c by "no valid user ID found".
* Add remarks to the man page about OpenPGP OPTIONAL algorithms and
issue warnings when such an algorithm is used (disabled by expert modus?)
* See how we can handle the symlinks in ./gcrypt
* Return status information even when an imported key is already
stored. This is useful for GPGME's detailed info function.
* Add SIGSEGV handler to overcome zlib problems with truncated data.
* Describe some pitfalls when using EGD. Check that ~/.gnupg/entropy
really is the default. What about needed permission?
* Use --output for keylistings too.
* Using an expired key for signing should give an error message
"expired key" and not "unusable key'. Furthermore the error should
also be thrown when the default key has expired. Reported by
Eric.VanBuggenhaut@AdValvas.be.
* Never allocate packet memory with a m-alloc, but use a specific function.
* pause scrolling help in --edit-key and elsewhere.
* Should we change names like mpi_write in g10/ so that we don't
use the prefix mpi here?
* getkey does not return revoked/expired keys - therefore it is not
possible to override it.
* Implement the AXP syscall to enable bus traps for GLIB 2
* Selection using +wordlist does not work.
What about adding a feature -word to the +wordlist search mode.
* parse a paramter file to do automatic key generation and to set
parameters for the keygen menu. Far easier than to do a dialog for
each required parameter.
* Check the changes to the gpg random gatherer on all W32 platforms.
* Put a note into readme.w32 that there is a man page and a options
file; write the registry stuff in regedit format.
* Show more info does not work from edit->trust
* set default charset from nl_langinfo.
* Check that no secret temporary results are stored in the result parameter
of the mpi functions. We have already done this for mpi-mul.c
* We need another special packet at the end of a clearsign message to mark
it's end and allow for multiple signature for one message. And
add a real grammar to the code in mainproc.c
* Option to warn when a non MDC message is decrypted?
* If there is no secure memory, allocate more memory for the secure
memory block or do it in all cases.
* add some minor things vor VMS.
* Use DSA keys with the test suite (partly done)
* Fix the bug in the mips assembler code
* Add a way to show the fingerprint of an key signator's keys
* Add an is_valid flag to each user ID.
* Replace the printing of the user name by [self-signature] when
appropriate so that a key listing does not get clobbered.
* Concatenated encryption messages don't work corectly - only the
first one is processed.
* Add option to put the list of recipients (from the encryption
layer) into the signatures notation data.
* With option -i prompt before adding a key to the keyring and show some
info what we are about to add.
* Speed up calculation of key validation.
* --disable-asm should still assemble _udiv_qrnnd when needed
* Skip RO keyrings when importing a key.
* Get new assembler stuff from gmp 3.1
* Use the newest encryption key if only the main key has been given.
* replace the keyserver stuff either by a call to a specialized
utility and SOCKSify this utility.
* Check the beginning of file to detect already compressed files (gzip,
bzip2, xdelta and some picture formats)
* Delay the read of the passphrase-fd afte a NEED_PASSPHRASE. But this
may break some scripts.
* as soon as we have moved to KBX, we can check signatures at all places
because there is no perfomance drawback as we can store the result of
a verification in the KBX. This enable us to better print information on
revoked user IDs and signatures. Well, caching of non-self-signatures
will still be complicated.
Nice to have
------------
* Official test vectors for 3DES-EDE3
* Use new-format headers for compressed packets.
The advantage is that a garbled zip files can be better detected.
* use DEL and ^H for erasing the previous character (util/ttyio.c).
or better readline.
* Print a warning if the directory mode is wrong.
* Do a real fix for bug #7 or document that it is a PGP 5 error.
* preferences of hash algorithms are not yet used.
* Replace the SIGUSR1 stuff by semaphores to avoid loss of a signal.
or use POSIX.4 realtime signals. Overhaul the interface and the
test program. Use it with the test suite?
* add test cases for invalid data (scrambled armor or other random data)
* add checking of armor trailers
* Burn the buffers used by fopen(), or use read(2). Does this
really make sense? And while we are at it: implement a secure deletion
stuff?
* add checking of armor trailers. Try to detect garbled header lines.
* the pubkey encrypt functions should do some sanity checks.
* dynload: implement the hint stuff.
* "gpg filename.tar.gz.asc" should work like --verify (-sab).
* for messages created with "-t", it might make sense to append the
verification status of the message to the output (i.e. write something to
the --output file and not only to stderr.
* configure option where to find zlib
* Display more validity information about the user IDs at certain places.
We need a more general function to extract such kind of info from the
trustdb.
* Evaluate whether it make sense to replace the namehashs either by
using the user ID directly or by using pointers into the trustdb.
* --gen-prime may trigger a log_bug; should be a log_fatal.
the --output file and not only to stderr. However the problem is
that we consider the message transpatrent and don't have any
indication of the used character set. To implement this feature
we need to make sure that all output is plain 7 bit ascii but
given that we need to print a user name, this does not make sense
at all. The only way this can be implemented is by assuming that
the message is encoded in utf8 and hope tht everyone starts to use
utf8 instead of latin-1 or whatever RSN. Hmmm, I myself should
start with this.
* keyflags don't distinguish between {certify,signature}-only.
* Instead of issuing a "signature packet without keyid" gpg should
try to get the keyID from a corresponding one-pass signature
packet (See bug report 817). This is not easy to do as we don't
store the one-pass packets.
* cat foo | gpg --sign | gpg --list-packets
Does not list the signature packet.
* Replace wording "It is NOT certain that the key belongs to its owner..."
with "It is NOT certain that the key belongs to the person named in the
user ID..." (or similar but clearer message).
Things we won't do
------------------
* New option --file-remove path-to-wipe-program ?

116
acconfig.h
View File

@ -1,116 +0,0 @@
/* acconfig.h - used by autoheader to make config.h.in
* Copyright (C) 1998 Free Software Foundation, Inc.
*
* This file is part of GNUPG.
*
* GNUPG is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* GNUPG is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA
*/
#ifndef GNUPG_CONFIG_H
#define GNUPG_CONFIG_H
/* need this, because some autoconf tests rely on this (e.g. stpcpy)
* and it should be used for new programs
*/
#define _GNU_SOURCE 1
@TOP@
#undef M_DEBUG
#undef M_GUARD
#undef VERSION
#undef PACKAGE
#undef PRINTABLE_OS_NAME
#undef IS_DEVELOPMENT_VERSION
/* Define if your locale.h file contains LC_MESSAGES. */
#undef HAVE_LC_MESSAGES
/* Define to 1 if NLS is requested. */
#undef ENABLE_NLS
/* Define as 1 if you have catgets and don't want to use GNU gettext. */
#undef HAVE_CATGETS
/* Define as 1 if you have gettext and don't want to use GNU gettext. */
#undef HAVE_GETTEXT
/* libintl.h is available; this is obsolete because if we don't have
* this header we use a symlink to the one in intl/ */
#undef HAVE_LIBINTL_H
#undef HAVE_STPCPY
#undef HAVE_MLOCK
#undef BIG_ENDIAN_HOST
#undef LITTLE_ENDIAN_HOST
#undef HAVE_BYTE_TYPEDEF
#undef HAVE_USHORT_TYPEDEF
#undef HAVE_ULONG_TYPEDEF
#undef HAVE_U16_TYPEDEF
#undef HAVE_U32_TYPEDEF
#undef HAVE_BROKEN_MLOCK
/* defined if we have a /dev/random and /dev/urandom */
#undef HAVE_DEV_RANDOM
/* and the real names of the random devices */
#undef NAME_OF_DEV_RANDOM
#undef NAME_OF_DEV_URANDOM
/* Linux has an ioctl */
#undef HAVE_DEV_RANDOM_IOCTL
/* see cipher/rndegd.c */
#undef EGD_SOCKET_NAME
#undef USE_DYNAMIC_LINKING
#undef HAVE_DL_DLOPEN
#undef HAVE_DL_SHL_LOAD
#undef HAVE_DLD_DLD_LINK
#undef USE_SHM_COPROCESSING
#undef IPC_HAVE_SHM_LOCK
#undef IPC_RMID_DEFERRED_RELEASE
/* set this to limit filenames to the 8.3 format */
#undef USE_ONLY_8DOT3
/* defined if we must run on a stupid file system */
#undef HAVE_DRIVE_LETTERS
/* defined if we run on some of the PCDOS like systems (DOS, Windoze. OS/2)
* with special properties like no file modes */
#undef HAVE_DOSISH_SYSTEM
/* because the Unix gettext has to much overhead on MingW32 systems
* and these systems lack Posix functions, we use a simplified version
* of gettext */
#undef USE_SIMPLE_GETTEXT
/* At some point in the system we need to know that we use the Windows
* random module. */
#undef USE_STATIC_RNDW32
#undef USE_CAPABILITIES
/* Some systems have mkdir that takes a single argument. */
#undef MKDIR_TAKES_ONE_ARG
@BOTTOM@
#include "gnupg-defs.h"
#endif /*GNUPG_CONFIG_H*/

247
acinclude.m4
View File

@ -1,5 +1,21 @@
dnl macros to configure g10
dnl macros to configure gnupg
dnl Copyright (C) 1998, 1999, 2000, 2001 Free Software Foundation, Inc.
dnl
dnl This file is part of GnuPG.
dnl
dnl GnuPG is free software; you can redistribute it and/or modify
dnl it under the terms of the GNU General Public License as published by
dnl the Free Software Foundation; either version 2 of the License, or
dnl (at your option) any later version.
dnl
dnl GnuPG is distributed in the hope that it will be useful,
dnl but WITHOUT ANY WARRANTY; without even the implied warranty of
dnl MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
dnl GNU General Public License for more details.
dnl
dnl You should have received a copy of the GNU General Public License
dnl along with this program; if not, write to the Free Software
dnl Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA
dnl GNUPG_MSG_PRINT(STRING)
dnl print a message
@ -15,37 +31,15 @@ dnl
AC_DEFUN(GNUPG_CHECK_TYPEDEF,
[ AC_MSG_CHECKING(for $1 typedef)
AC_CACHE_VAL(gnupg_cv_typedef_$1,
[AC_TRY_COMPILE([#include <stdlib.h>
[AC_TRY_COMPILE([#define _GNU_SOURCE 1
#include <stdlib.h>
#include <sys/types.h>], [
#undef $1
int a = sizeof($1);
], gnupg_cv_typedef_$1=yes, gnupg_cv_typedef_$1=no )])
AC_MSG_RESULT($gnupg_cv_typedef_$1)
if test "$gnupg_cv_typedef_$1" = yes; then
AC_DEFINE($2)
fi
])
dnl GNUPG_FIX_HDR_VERSION(FILE, NAME)
dnl Make the version number in gcrypt/gcrypt.h the same as the one here.
dnl (this is easier than to have a .in file just for one substitution)
dnl
AC_DEFUN(GNUPG_FIX_HDR_VERSION,
[ sed "s/^#define $2 \".*/#define $2 \"$VERSION\"/" $srcdir/$1 > $srcdir/$1.tmp
if cmp -s $srcdir/$1 $srcdir/$1.tmp 2>/dev/null; then
rm -f $srcdir/$1.tmp
else
rm -f $srcdir/$1
if mv $srcdir/$1.tmp $srcdir/$1 ; then
:
else
AC_MSG_ERROR([[
*** Failed to fix the version string macro $2 in $1.
*** The old file has been saved as $1.tmp
]])
fi
AC_MSG_WARN([fixed the $2 macro in $1])
AC_DEFINE($2,1,[Defined if a `]$1[' is typedef'd])
fi
])
@ -53,7 +47,7 @@ AC_DEFUN(GNUPG_FIX_HDR_VERSION,
dnl GNUPG_CHECK_GNUMAKE
dnl
AC_DEFUN(GNUPG_CHECK_GNUMAKE,
[
[
if ${MAKE-make} --version 2>/dev/null | grep '^GNU ' >/dev/null 2>&1; then
:
else
@ -67,6 +61,7 @@ AC_DEFUN(GNUPG_CHECK_GNUMAKE,
fi
])
dnl GNUPG_CHECK_FAQPROG
dnl
AC_DEFUN(GNUPG_CHECK_FAQPROG,
@ -82,39 +77,33 @@ AC_DEFUN(GNUPG_CHECK_FAQPROG,
AC_SUBST(FAQPROG)
AM_CONDITIONAL(WORKING_FAQPROG, test "$working_faqprog" = "yes" )
if test $working_faqprog = no; then
AC_MSG_WARN([[
***
*** It seems that the faqprog.pl program is not installed.
*** Unless you do not change the source of the FAQs it is not required.
*** The working version of this utility should be available at:
*** ftp://ftp.gnupg.org/pub/gcrypt/contrib/faqprog.pl
***]])
fi
])
dnl if test $working_faqprog = no; then
dnl AC_MSG_WARN([[
dnl ***
dnl *** It seems that the faqprog.pl program is not installed;
dnl *** however it is only needed if you want to change the FAQ.
dnl *** (faqprog.pl should be available at:
dnl *** ftp://ftp.gnupg.org/pub/gcrypt/contrib/faqprog.pl )
dnl *** No need to worry about this warning.
dnl ***]])
dnl fi
])
dnl GNUPG_LINK_FILES( SRC, DEST )
dnl same as AC_LINK_FILES, but collect the files to link in
dnl some special variables and do the link
dnl when GNUPG_DO_LINK_FILES is called
dnl This is a workaround for AC_LINK_FILES, because it does not work
dnl correct when using a caching scheme
dnl GNUPG_CHECK_DOCBOOK_TO_TEXI
dnl
define(GNUPG_LINK_FILES,
[ if test "x$wk_link_files_src" = "x"; then
wk_link_files_src="$1"
wk_link_files_dst="$2"
else
wk_link_files_src="$wk_link_files_src $1"
wk_link_files_dst="$wk_link_files_dst $2"
AC_DEFUN(GNUPG_CHECK_DOCBOOK_TO_TEXI,
[
AC_CHECK_PROG(DOCBOOK_TO_TEXI, docbook2texi, yes, no)
AC_MSG_CHECKING(for sgml to texi tools)
working_sgmltotexi=no
if test "$ac_cv_prog_DOCBOOK_TO_TEXI" = yes; then
if sgml2xml -v /dev/null 2>&1 | grep 'SP version' >/dev/null 2>&1 ; then
working_sgmltotexi=yes
fi
fi
])
define(GNUPG_DO_LINK_FILES,
[ AC_LINK_FILES( $wk_link_files_src, $wk_link_files_dst )
])
AC_MSG_RESULT($working_sgmltotexi)
AM_CONDITIONAL(HAVE_DOCBOOK_TO_TEXI, test "$working_sgmltotexi" = "yes" )
])
@ -158,9 +147,11 @@ define(GNUPG_CHECK_ENDIAN,
])
AC_MSG_RESULT([$gnupg_cv_c_endian])
if test "$gnupg_cv_c_endian" = little; then
AC_DEFINE(LITTLE_ENDIAN_HOST)
AC_DEFINE(LITTLE_ENDIAN_HOST,1,
[Defined if the host has little endian byte ordering])
else
AC_DEFINE(BIG_ENDIAN_HOST)
AC_DEFINE(BIG_ENDIAN_HOST,1,
[Defined if the host has big endian byte ordering])
fi
])
@ -328,7 +319,8 @@ define(GNUPG_CHECK_IPC,
gnupg_cv_ipc_rmid_deferred_release="assume-no")
)
if test "$gnupg_cv_ipc_rmid_deferred_release" = "yes"; then
AC_DEFINE(IPC_RMID_DEFERRED_RELEASE)
AC_DEFINE(IPC_RMID_DEFERRED_RELEASE,1,
[Defined if we can do a deferred shm release])
AC_MSG_RESULT(yes)
else
if test "$gnupg_cv_ipc_rmid_deferred_release" = "no"; then
@ -351,7 +343,8 @@ define(GNUPG_CHECK_IPC,
)
)
if test "$gnupg_cv_ipc_have_shm_lock" = "yes"; then
AC_DEFINE(IPC_HAVE_SHM_LOCK)
AC_DEFINE(IPC_HAVE_SHM_LOCK,1,
[Defined if a SysV shared memory supports the LOCK flag])
AC_MSG_RESULT(yes)
else
AC_MSG_RESULT(no)
@ -384,7 +377,7 @@ define(GNUPG_CHECK_MLOCK,
#endif
], [
int i;
/* glibc defines this for functions which it implements
* to always fail with ENOSYS. Some functions are actually
* named something starting with __ and the normal name
@ -399,7 +392,8 @@ define(GNUPG_CHECK_MLOCK,
gnupg_cv_mlock_is_in_sys_mman=yes,
gnupg_cv_mlock_is_in_sys_mman=no)])
if test "$gnupg_cv_mlock_is_in_sys_mman" = "yes"; then
AC_DEFINE(HAVE_MLOCK)
AC_DEFINE(HAVE_MLOCK,1,
[Defined if the system supports an mlock() call])
fi
fi
fi
@ -439,8 +433,10 @@ define(GNUPG_CHECK_MLOCK,
)
)
if test "$gnupg_cv_have_broken_mlock" = "yes"; then
AC_DEFINE(HAVE_BROKEN_MLOCK)
AC_DEFINE(HAVE_BROKEN_MLOCK,1,
[Defined if the mlock() call does not work])
AC_MSG_RESULT(yes)
AC_CHECK_FUNCS(plock)
else
if test "$gnupg_cv_have_broken_mlock" = "no"; then
AC_MSG_RESULT(no)
@ -452,10 +448,8 @@ define(GNUPG_CHECK_MLOCK,
])
################################################################
# GNUPG_PROG_NM - find the path to a BSD-compatible name lister
################################################################
AC_DEFUN(GNUPG_PROG_NM,
[AC_MSG_CHECKING([for BSD-compatible nm])
AC_CACHE_VAL(ac_cv_path_NM,
@ -667,7 +661,7 @@ AC_CHECK_TOOL(AS, as, false)
AC_DEFUN(GNUPG_SYS_SYMBOL_UNDERSCORE,
[tmp_do_check="no"
case "${target}" in
i386-emx-os2 | i[3456]86-pc-os2*emx | i386-pc-msdosdjgpp)
i386-emx-os2 | i[3456]86-pc-os2*emx | i386-pc-msdosdjgpp | *-*-cygwin)
ac_cv_sys_symbol_underscore=yes
;;
*)
@ -718,7 +712,7 @@ fi
AC_MSG_RESULT($ac_cv_sys_symbol_underscore)
if test x$ac_cv_sys_symbol_underscore = xyes; then
AC_DEFINE(WITH_SYMBOL_UNDERSCORE,1,
[define if compiled symbols have a leading underscore])
[Defined if compiled symbols have a leading underscore])
fi
])
@ -741,119 +735,8 @@ AC_CACHE_CHECK([if mkdir takes one argument], gnupg_cv_mkdir_takes_one_arg,
#endif], [mkdir ("foo", 0);],
gnupg_cv_mkdir_takes_one_arg=no, gnupg_cv_mkdir_takes_one_arg=yes)])
if test $gnupg_cv_mkdir_takes_one_arg = yes ; then
AC_DEFINE(MKDIR_TAKES_ONE_ARG)
AC_DEFINE(MKDIR_TAKES_ONE_ARG,1,
[Defined if mkdir() does not take permission flags])
fi
])
dnl GPH_PROG_DOCBOOK()
dnl Check whether we have the needed Docbook environment
dnl and issue a warning if this is not the case.
dnl
dnl This test defines these variables for substitution:
dnl DB2HTML - command used to convert Docbook to HTML
dnl DB2TEX - command used to convert Docbook to TeX
dnl DB2MAN - command used to convert Docbook to man pages
dnl JADE - command to invoke jade
dnl JADETEX - command to invoke jadetex
dnl DSL_FOR_HTML - the stylesheet used to for the Docbook->HTML conversion
dnl The following make conditionals are defined
dnl HAVE_DB2MAN - defined when db2man is available
dnl HAVE_DB2TEX - defined when db2tex is available
dnl HAVE_DB2HTML - defined when db2html is available
dnl HAVE_DOCBOOK - defined when the entire Docbook environment is present
dnl HAVE_JADE - defined when jade is installed
dnl HAVE_JADETEX - defined when jadetex is installed
dnl
dnl (wk 2000-02-17)
dnl
AC_DEFUN(GPH_PROG_DOCBOOK,
[ AC_REQUIRE([AC_CONFIG_AUX_DIR_DEFAULT])dnl
all=yes
AC_PATH_PROG(DB2MAN, docbook-to-man, no)
test "$DB2MAN" = no && all=no
AM_CONDITIONAL(HAVE_DB2MAN, test "$DB2MAN" != no )
AC_PATH_PROG(JADE, jade, no)
test "$JADE" = no && all=no
AM_CONDITIONAL(HAVE_JADE, test "$JADE" != no )
AC_PATH_PROG(JADETEX, jadetex, no)
test "$JADETEX" = no && all=no
AM_CONDITIONAL(HAVE_JADETEX, test "$JADETEX" != no )
stylesheet_dirs='
/usr/local/lib/dsssl/stylesheets/docbook
/usr/local/share/dsssl/stylesheets/docbook
/usr/local/lib/sgml/stylesheet/dsssl/docbook/nwalsh
/usr/local/share/sgml/stylesheet/dsssl/docbook/nwalsh
/usr/lib/dsssl/stylesheets/docbook
/usr/share/dsssl/stylesheets/docbook
/usr/lib/sgml/stylesheet/dsssl/docbook/nwalsh
/usr/share/sgml/stylesheet/dsssl/docbook/nwalsh
'
AC_MSG_CHECKING(for TeX stylesheet)
dsl=none
for d in ${stylesheet_dirs}; do
file=${d}/print/docbook.dsl
if test -f $file; then
dsl=$file
break
fi
done
AC_MSG_RESULT([$dsl])
okay=no
if test $dsl = none ; then
DB2TEX="$missing_dir/missing db2tex"
all=no
else
DB2TEX="$JADE -t tex -i tex -d $dsl"
okay=yes
fi
AC_SUBST(DB2TEX)
AM_CONDITIONAL(HAVE_DB2TEX, test $okay = yes )
if ( $ac_aux_dir/db2html.in --version) < /dev/null > /dev/null 2>&1; then
:
else
AC_ERROR([needed $ac_aux_dir/db2html.in not found])
fi
AC_MSG_CHECKING(for HTML stylesheet)
DSL_FOR_HTML="none"
for d in ${stylesheet_dirs}; do
file=${d}/html/docbook.dsl
if test -f $file; then
DSL_FOR_HTML=$file
break
fi
done
AC_MSG_RESULT([$DSL_FOR_HTML])
okay=no
if test $DSL_FOR_HTML = none ; then
DB2HTML="$missing_dir/missing db2html"
all=no
else
DB2HTML="`cd $ac_aux_dir && pwd`/db2html --copyfiles"
okay=yes
fi
AC_SUBST(DB2HTML)
AC_SUBST(DSL_FOR_HTML)
AM_CONDITIONAL(HAVE_DB2HTML, test $okay = yes )
AM_CONDITIONAL(HAVE_DOCBOOK, test "$all" != yes )
if test $all = no ; then
AC_MSG_WARN([[
***
*** It seems that the Docbook environment is not installed as required.
*** We will try to build everything, but if you either touch some files
*** or use a bogus make tool, you may run into problems.
*** Docbook is normally only needed to build the documentation.
***]])
fi
])
dnl *-*wedit:notab*-* Please keep this as the last line.

5
autogen.sh Executable file
View File

@ -0,0 +1,5 @@
#!/bin/sh
autogen_sh="`dirname $0`/scripts/autogen.sh"
exec $autogen_sh $*

994
configure.ac Normal file
View File

@ -0,0 +1,994 @@
dnl Configure.ac script for GnuPG
dnl Copyright (C) 1998, 1999, 2000, 2001, 2002 Free Software Foundation, Inc.
dnl
dnl This file is part of GnuPG.
dnl
dnl GnuPG is free software; you can redistribute it and/or modify
dnl it under the terms of the GNU General Public License as published by
dnl the Free Software Foundation; either version 2 of the License, or
dnl (at your option) any later version.
dnl
dnl GnuPG is distributed in the hope that it will be useful,
dnl but WITHOUT ANY WARRANTY; without even the implied warranty of
dnl MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
dnl GNU General Public License for more details.
dnl
dnl You should have received a copy of the GNU General Public License
dnl along with this program; if not, write to the Free Software
dnl Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA
dnl
dnl (Process this file with autoconf to produce a configure script.)
dnlAC_REVISION($Revision$)dnl
AC_PREREQ(2.52)
AC_INIT(gnupg, 1.1.90, bug-gnupg@gnu.org)
ALL_LINGUAS="cs da de eo el es et fr gl id it ja nl pl pt_BR pt sv tr"
static_modules="sha1 md5 rmd160"
static_random_module=""
PACKAGE=$PACKAGE_NAME
VERSION=$PACKAGE_VERSION
AC_CONFIG_AUX_DIR(scripts)
AC_CONFIG_SRCDIR(g10/g10.c)
AC_CANONICAL_TARGET()
AM_INIT_AUTOMAKE($PACKAGE, $VERSION)
AM_CONFIG_HEADER(config.h)
AC_SUBST(PACKAGE)
AC_SUBST(VERSION)
AC_DEFINE_UNQUOTED(PACKAGE, "$PACKAGE", [Name of this package])
AC_DEFINE_UNQUOTED(VERSION, "$VERSION", [Version of this package])
AC_DEFINE(_GNU_SOURCE,1,[Some tests rely on this (stpcpy) and it should be used for new programs anyway])
AC_PROG_AWK
MODULES_IN_CIPHER=`$AWK '/^EXTRA_PROGRAMS/ { for(i=3;i<=NF;i++) print $i}' \
$srcdir/cipher/Makefile.am`
dnl
dnl Check for random module options
dnl
dnl Fixme: get the list of available modules from MODULES_IN_CIPHER
dnl and check against this list
AC_MSG_CHECKING([which static random module to use])
AC_ARG_ENABLE(static-rnd,
[ --enable-static-rnd=[egd|unix|linux|none] ],
[use_static_rnd=$enableval], [use_static_rnd=default] )
if test "$use_static_rnd" = no; then
use_static_rnd=default
fi
case "$use_static_rnd" in
egd | linux | unix | none | default )
AC_MSG_RESULT($use_static_rnd)
;;
* )
AC_MSG_RESULT(invalid argument)
AC_MSG_ERROR(there is no random module rnd$use_static_rnd)
;;
esac
AC_ARG_WITH(egd-socket,
[ --with-egd-socket=NAME Use NAME for the EGD socket)],
egd_socket_name="$withval", egd_socket_name="" )
AC_DEFINE_UNQUOTED(EGD_SOCKET_NAME, "$egd_socket_name",
[Define if you don't want the default EGD socket name.
For details see cipher/rndegd.c])
dnl
dnl See whether the user wants to disable checking for /dev/random
AC_MSG_CHECKING([whether use of /dev/random is requested])
AC_ARG_ENABLE(dev-random,
[ --disable-dev-random disable the use of dev random],
try_dev_random=$enableval, try_dev_random=yes)
AC_MSG_RESULT($try_dev_random)
dnl
dnl Check other options
dnl
AC_MSG_CHECKING([whether use of extensions is requested])
AC_ARG_ENABLE(dynload,
[ --disable-dynload disable use of extensions],
try_dynload=$enableval, try_dynload=yes)
AC_MSG_RESULT($try_dynload)
AC_MSG_CHECKING([whether assembler modules are requested])
AC_ARG_ENABLE(asm,
[ --disable-asm do not use assembler modules],
try_asm_modules=$enableval, try_asm_modules=yes)
AC_MSG_RESULT($try_asm_modules)
AC_MSG_CHECKING([whether memory guard is requested])
AC_ARG_ENABLE(m-guard,
[ --enable-m-guard enable memory guard facility],
use_m_guard=$enableval, use_m_guard=no)
AC_MSG_RESULT($use_m_guard)
if test "$use_m_guard" = yes ; then
AC_DEFINE(M_GUARD,1,[Define to use the (obsolete) malloc guarding feature])
fi
AC_MSG_CHECKING([whether to enable external program execution])
AC_ARG_ENABLE(exec,
[ --disable-exec disable external program execution],
use_exec=$enableval, use_exec=yes)
AC_MSG_RESULT($use_exec)
if test "$use_exec" = no ; then
AC_DEFINE(NO_EXEC,1,[Define to disable external program execution])
fi
if test "$use_exec" = yes ; then
AC_MSG_CHECKING([whether LDAP keyserver support is requested])
AC_ARG_ENABLE(ldap,
[ --disable-ldap disable LDAP keyserver interface],
try_ldap=$enableval, try_ldap=yes)
AC_MSG_RESULT($try_ldap)
AC_MSG_CHECKING([whether experimental external hkp keyserver support is requested])
AC_ARG_ENABLE(external-hkp,
[ --enable-external-hkp enable experimental external HKP keyserver interface],
try_hkp=$enableval, try_hkp=no)
AC_MSG_RESULT($try_hkp)
if test "$try_hkp" = yes ; then
GPGKEYS_HKP="gpgkeys_hkp"
AC_DEFINE(USE_EXTERNAL_HKP,1,[define to use the experimental external HKP keyserver interface])
fi
AC_MSG_CHECKING([whether email keyserver support is requested])
AC_ARG_ENABLE(mailto,
[ --disable-mailto disable email keyserver interface],
try_mailto=$enableval, try_mailto=yes)
AC_MSG_RESULT($try_mailto)
fi
AC_MSG_CHECKING([whether included zlib is requested])
AC_ARG_WITH(included-zlib,
[ --with-included-zlib use the zlib code included here],
[g10_force_zlib=yes], [g10_force_zlib=no] )
AC_MSG_RESULT($g10_force_zlib)
dnl
dnl Check whether we want to use Linux capabilities
dnl
AC_MSG_CHECKING([whether use of capabilities is requested])
AC_ARG_WITH(capabilities,
[ --with-capabilities use linux capabilities [default=no]],
[use_capabilities="$withval"],[use_capabilities=no])
AC_MSG_RESULT($use_capabilities)
AH_BOTTOM([
#if !(defined(HAVE_FORK) && defined(HAVE_PIPE) && defined(HAVE_WAITPID))
#define EXEC_TEMPFILE_ONLY
#endif
#include "g10defs.h"
])
AM_MAINTAINER_MODE
dnl Checks for programs.
dnl
dnl Setup some stuff depending on host/target.
dnl
case "${target}" in
*-*-mingw32*)
disallowed_modules="rndunix rndlinux rndegd"
;;
*-*-cygwin*)
disallowed_modules="rndunix rndlinux rndegd tiger"
;;
*)
disallowed_modules="rndw32"
;;
esac
AC_PROG_MAKE_SET
AM_SANITY_CHECK
missing_dir=`cd $ac_aux_dir && pwd`
AM_MISSING_PROG(ACLOCAL, aclocal, $missing_dir)
AM_MISSING_PROG(AUTOCONF, autoconf, $missing_dir)
AM_MISSING_PROG(AUTOMAKE, automake, $missing_dir)
AM_MISSING_PROG(AUTOHEADER, autoheader, $missing_dir)
AM_MISSING_PROG(MAKEINFO, makeinfo, $missing_dir)
AC_PROG_CC
AC_PROG_CPP
AC_PATH_PROG(PERL,"perl")
AC_ISC_POSIX
AC_SYS_LARGEFILE
AC_PROG_INSTALL
AC_PROG_AWK
AC_CHECK_PROG(DOCBOOK_TO_MAN, docbook-to-man, yes, no)
AM_CONDITIONAL(HAVE_DOCBOOK_TO_MAN, test "$ac_cv_prog_DOCBOOK_TO_MAN" = yes)
GNUPG_CHECK_FAQPROG
GNUPG_CHECK_DOCBOOK_TO_TEXI
MPI_OPT_FLAGS=""
try_gettext=yes
case "${target}" in
*-*-mingw32*|*-*-cygwin*)
# special stuff for Windoze NT
ac_cv_have_dev_random=no
AC_DEFINE(USE_ONLY_8DOT3,1,
[set this to limit filenames to the 8.3 format])
AC_DEFINE(HAVE_DRIVE_LETTERS,1,
[defined if we must run on a stupid file system])
AC_DEFINE(HAVE_DOSISH_SYSTEM,1,
[defined if we run on some of the PCDOS like systems
(DOS, Windoze. OS/2) with special properties like
no file modes])
AC_DEFINE(USE_SIMPLE_GETTEXT,1,
[because the Unix gettext has too much overhead on
MingW32 systems and these systems lack Posix functions,
we use a simplified version of gettext])
try_gettext="no"
;;
i?86-emx-os2 | i?86-*-os2*emx )
# OS/2 with the EMX environment
ac_cv_have_dev_random=no
AC_DEFINE(HAVE_DRIVE_LETTERS)
AC_DEFINE(HAVE_DOSISH_SYSTEM)
try_gettext="no"
;;
i?86-*-msdosdjgpp*)
# DOS with the DJGPP environment
ac_cv_have_dev_random=no
AC_DEFINE(HAVE_DRIVE_LETTERS)
AC_DEFINE(HAVE_DOSISH_SYSTEM)
try_gettext="no"
;;
*-*-freebsd*)
# FreeBSD
CPPFLAGS="$CPPFLAGS -I/usr/local/include"
LDFLAGS="$LDFLAGS -L/usr/local/lib"
;;
*-*-hpux*)
if test -z "$GCC" ; then
CFLAGS="$CFLAGS -Ae -D_HPUX_SOURCE"
fi
;;
*-dec-osf4*)
if test -z "$GCC" ; then
# Suppress all warnings
# to get rid of the unsigned/signed char mismatch warnings.
CFLAGS="$CFLAGS -w"
fi
;;
*-dec-osf5*)
if test -z "$GCC" ; then
# Use the newer compiler `-msg_disable ptrmismatch' to
# get rid of the unsigned/signed char mismatch warnings.
# Using this may hide other pointer mismatch warnings, but
# it at least lets other warning classes through
CFLAGS="$CFLAGS -msg_disable ptrmismatch"
fi
;;
m68k-atari-mint)
;;
*)
;;
esac
AC_SUBST(MPI_OPT_FLAGS)
GNUPG_SYS_SYMBOL_UNDERSCORE
GNUPG_CHECK_PIC
GNUPG_CHECK_EXPORTDYNAMIC
if test "$NO_PIC" = yes; then
try_dynload=no
fi
dnl Must check for network library requirements before doing link tests
dnl for ldap, for example. If ldap libs are static (or dynamic and without
dnl ELF runtime link paths), then link will fail and LDAP support won't
dnl be detected.
AC_CHECK_FUNC(gethostbyname, , AC_CHECK_LIB(nsl, gethostbyname,
[NETLIBS="-lnsl $NETLIBS"]))
AC_CHECK_FUNC(setsockopt, , AC_CHECK_LIB(socket, setsockopt,
[NETLIBS="-lsocket $NETLIBS"]))
# Try and link a LDAP test program to weed out unusable LDAP
# libraries. -lldap -llber -lresolv is for OpenLDAP. -lldapssl41 is
# for Mozilla LDAP.
if test "$try_ldap" = yes ; then
for LDAPLIBS in "-lldap -llber -lresolv" "-lldapssl41"; do
_ldap_save_libs=$LIBS
LIBS="$LDAPLIBS $NETLIBS $LIBS"
AC_MSG_CHECKING([whether LDAP via \"$LDAPLIBS\" is present and sane])
AC_TRY_LINK([#include <ldap.h>],[ldap_open("foobar",1234);],
[gnupg_cv_func_ldap_init=yes],[gnupg_cv_func_ldap_init=no])
AC_MSG_RESULT([$gnupg_cv_func_ldap_init])
if test $gnupg_cv_func_ldap_init = no; then
AC_MSG_CHECKING([whether I can make LDAP be sane with lber.h])
AC_TRY_LINK([#include <lber.h>
#include <ldap.h>],[ldap_open("foobar",1234);],
[gnupg_cv_func_ldaplber_init=yes],[gnupg_cv_func_ldaplber_init=no])
AC_MSG_RESULT([$gnupg_cv_func_ldaplber_init])
fi
if test "$gnupg_cv_func_ldaplber_init" = yes ; then
AC_DEFINE(NEED_LBER_H,1,[Define if the LDAP library requires including lber.h before ldap.h])
fi
if test "$gnupg_cv_func_ldap_init" = yes || \
test "$gnupg_cv_func_ldaplber_init" = yes ; then
GPGKEYS_LDAP="gpgkeys_ldap"
fi
LIBS=$_ldap_save_libs
if test "$GPGKEYS_LDAP" != "" ; then break; fi
done
fi
AC_SUBST(GPGKEYS_LDAP)
AC_SUBST(LDAPLIBS)
if test "$try_mailto" = yes ; then
AC_PATH_PROG(SENDMAIL,sendmail,,$PATH:/usr/sbin:/usr/libexec:/usr/lib)
if test "$ac_cv_path_SENDMAIL" ; then
GPGKEYS_MAILTO="gpgkeys_mailto"
fi
fi
AC_SUBST(GPGKEYS_MAILTO)
AC_SUBST(GPGKEYS_HKP)
case "${target}" in
*-*-mingw32*)
PRINTABLE_OS_NAME="MingW32"
;;
*-*-cygwin*)
PRINTABLE_OS_NAME="Cygwin32"
;;
i?86-emx-os2 | i?86-*-os2*emx )
PRINTABLE_OS_NAME="OS/2"
;;
i?86-*-msdosdjgpp*)
PRINTABLE_OS_NAME="MSDOS/DJGPP"
try_dynload=no
;;
*-linux*)
PRINTABLE_OS_NAME="GNU/Linux"
;;
dnl let that after linux to avoid gnu-linux problems
*-gnu*)
PRINTABLE_OS_NAME="GNU/Hurd"
;;
*)
PRINTABLE_OS_NAME=`uname -s || echo "Unknown"`
;;
esac
AC_DEFINE_UNQUOTED(PRINTABLE_OS_NAME, "$PRINTABLE_OS_NAME",
[A human readable text with the name of the OS])
#
# 1. Set flags to be used for the extension modules
# 2. Set names of random devices
#
NAME_OF_DEV_RANDOM="/dev/random"
NAME_OF_DEV_URANDOM="/dev/urandom"
case "${target}" in
*-openbsd* | *-netbsd*)
NAME_OF_DEV_RANDOM="/dev/srandom"
NAME_OF_DEV_URANDOM="/dev/urandom"
DYNLINK_MOD_CFLAGS="-shared -rdynamic $CFLAGS_PIC -Wl,-Bshareable -Wl,-x"
;;
hpux*)
# if using the vendor (ANSI) compiler, arrange to have `-b' passed
# to the linker. If using gcc, it supports `-shared' to do the same.
if test -n "$GCC" ; then
DYNLINK_MOD_CFLAGS="-shared $CFLAGS_PIC"
else
DYNLINK_MOD_CFLAGS='-Wl,-b'
fi
;;
*-irix6.5*)
# Irix 6.5 (and probably a lot earlier, but I only still have
# access to 6.5.x) doesn't require any additional flags, as `-KPIC'
# is the default. Also, `-shared' works with the vendor compiler
DYNLINK_MOD_CFLAGS="-shared"
;;
alpha*-dec-osf*)
# osf (i.e. OSF/1, Digital UNIX, or Tru64 UNIX, pick any one...)
# on alpha doesn't require any PIC flags, everything is PIC.
# This may not be the case for osf ports to other machines, so
# hence the more specific match for target.
#
# Also, `-shared' works with the vendor compiler or gcc.
# -expect_unresolved turns off warnings about unresolved symbols.
DYNLINK_MOD_CFLAGS='-shared -Wl,-expect_unresolved,\*'
;;
*-solaris*)
# -shared is a gcc-ism. Find pic flags from GNUPG_CHECK_PIC.
if test -n "$GCC" ; then
DYNLINK_MOD_CFLAGS="-shared $CFLAGS_PIC"
else
DYNLINK_MOD_CFLAGS="-G $CFLAGS_PIC"
fi
;;
*-irix* | *-dec-osf* )
DYNLINK_MOD_CFLAGS="-shared $CFLAGS_PIC"
;;
*)
# -shared is a gcc-ism. Find pic flags from GNUPG_CHECK_PIC.
if test -n "$GCC" ; then
DYNLINK_MOD_CFLAGS="-shared $CFLAGS_PIC"
else
DYNLINK_MOD_CFLAGS="$CFLAGS_PIC"
fi
;;
esac
AC_DEFINE_UNQUOTED(NAME_OF_DEV_RANDOM, "$NAME_OF_DEV_RANDOM",
[defined to the name of the strong random device])
AC_DEFINE_UNQUOTED(NAME_OF_DEV_URANDOM, "$NAME_OF_DEV_URANDOM",
[defined to the name of the weaker random device])
AC_SUBST(MPI_OPT_FLAGS)
dnl Checks for libraries.
if test "$try_gettext" = yes; then
AM_GNU_GETTEXT
else
USE_NLS=no
USE_INCLUDED_LIBINTL=no
BUILD_INCLUDED_LIBINTL=no
AC_SUBST(USE_NLS)
AC_SUBST(USE_INCLUDED_LIBINTL)
AC_SUBST(BUILD_INCLUDED_LIBINTL)
fi
if test "$try_dynload" = yes ; then
AC_CHECK_LIB(dl,dlopen)
if test "$ac_cv_lib_dl_dlopen" = "yes"; then
AC_DEFINE(USE_DYNAMIC_LINKING,1,
[define to enable the use of extensions])
AC_DEFINE(HAVE_DL_DLOPEN,1,
[Defined when the dlopen function family is available])
DYNLINK_LDFLAGS="$CFLAGS_EXPORTDYNAMIC"
use_gnupg_extensions=yes
else
AC_CHECK_FUNCS(dlopen)
if test "$ac_cv_func_dlopen" = "yes"; then
AC_DEFINE(USE_DYNAMIC_LINKING)
AC_DEFINE(HAVE_DL_DLOPEN)
DYNLINK_LDFLAGS="$CFLAGS_EXPORTDYNAMIC"
use_gnupg_extensions=yes
else
AC_CHECK_LIB(dld,shl_load)
if test "$ac_cv_lib_dld_shl_load" = "yes"; then
AC_DEFINE(USE_DYNAMIC_LINKING)
AC_DEFINE(HAVE_DL_SHL_LOAD,1,
[Defined if the shl_load function family is available])
DYNLINK_LDFLAGS="$CFLAGS_EXPORTDYNAMIC"
use_gnupg_extensions=yes
dnl -----------------
dnl DLD is not ready for use. So I better disable this test
dnl
dnl AC_CHECK_LIB(dld,dld_link)
dnl if test "$ac_cv_lib_dld_dld_link" = "yes"; then
dnl AC_DEFINE(USE_DYNAMIC_LINKING)
dnl AC_DEFINE(HAVE_DLD_DLD_LINK)
dnl DYNLINK_LDFLAGS="$CFLAGS_EXPORTDYNAMIC"
dnl use_gnupg_extensions=yes
dnl ---------------
fi
fi
fi
else
AC_MSG_CHECKING(for dynamic loading)
DYNLINK_LDFLAGS=
DYNLINK_MOD_CFLAGS=
use_gnupg_extensions=no
AC_MSG_RESULT(has been disabled)
fi
AM_CONDITIONAL(ENABLE_GNUPG_EXTENSIONS, test "$use_gnupg_extensions" = yes )
AC_SUBST(DYNLINK_LDFLAGS)
AC_SUBST(DYNLINK_MOD_CFLAGS)
dnl Checks for header files.
AC_HEADER_STDC
AC_CHECK_HEADERS(unistd.h langinfo.h termio.h locale.h)
dnl Checks for typedefs, structures, and compiler characteristics.
AC_C_CONST
AC_C_INLINE
AC_TYPE_SIZE_T
AC_TYPE_SIGNAL
AC_DECL_SYS_SIGLIST
GNUPG_CHECK_ENDIAN
GNUPG_CHECK_TYPEDEF(byte, HAVE_BYTE_TYPEDEF)
GNUPG_CHECK_TYPEDEF(ushort, HAVE_USHORT_TYPEDEF)
GNUPG_CHECK_TYPEDEF(ulong, HAVE_ULONG_TYPEDEF)
GNUPG_CHECK_TYPEDEF(u16, HAVE_U16_TYPEDEF)
GNUPG_CHECK_TYPEDEF(u32, HAVE_U32_TYPEDEF)
AC_CHECK_SIZEOF(unsigned short, 2)
AC_CHECK_SIZEOF(unsigned int, 4)
AC_CHECK_SIZEOF(unsigned long, 4)
AC_CHECK_SIZEOF(unsigned long long, 0)
if test "$ac_cv_sizeof_unsigned_short" = "0" \
|| test "$ac_cv_sizeof_unsigned_int" = "0" \
|| test "$ac_cv_sizeof_unsigned_long" = "0"; then
AC_MSG_WARN([Hmmm, something is wrong with the sizes - using defaults]);
fi
dnl Checks for library functions.
AC_FUNC_FSEEKO
AC_FUNC_VPRINTF
AC_FUNC_FORK
AC_CHECK_FUNCS(strerror stpcpy strsep strlwr tcgetattr strtoul mmap)
AC_CHECK_FUNCS(strcasecmp strncasecmp)
AC_CHECK_FUNCS(memmove gettimeofday getrusage setrlimit clock_gettime)
AC_CHECK_FUNCS(atexit raise getpagesize strftime nl_langinfo setlocale)
AC_CHECK_FUNCS(waitpid wait4 sigaction sigprocmask rand pipe stat)
AC_REPLACE_FUNCS(mkdtemp)
#
# check for gethrtime and run a testprogram to see whether
# it is broken. It has been reported that some Solaris and HP UX systems
# raise an SIGILL
#
AC_CACHE_CHECK([for gethrtime],
[gnupg_cv_func_gethrtime],
[AC_TRY_LINK([#include <sys/times.h>],[
hrtime_t tv;
tv = gethrtime();
],
[gnupg_cv_func_gethrtime=yes],
[gnupg_cv_func_gethrtime=no])
])
if test $gnupg_cv_func_gethrtime = yes; then
AC_DEFINE([HAVE_GETHRTIME], 1,
[Define if you have the `gethrtime(2)' function.])
AC_CACHE_CHECK([whether gethrtime is broken],
[gnupg_cv_func_broken_gethrtime],
[AC_TRY_RUN([
#include <sys/times.h>
int main () {
hrtime_t tv;
tv = gethrtime();
}
],
[gnupg_cv_func_broken_gethrtime=no],
[gnupg_cv_func_broken_gethrtime=yes],
[gnupg_cv_func_broken_gethrtime=assume-no])
])
if test $gnupg_cv_func_broken_gethrtime = yes; then
AC_DEFINE([HAVE_BROKEN_GETHRTIME], 1,
[Define if `gethrtime(2)' does not work correctly i.e. issues a SIGILL.])
fi
fi
GNUPG_CHECK_MLOCK
GNUPG_FUNC_MKDIR_TAKES_ONE_ARG
dnl
dnl Check whether we can use Linux capabilities as requested
dnl
if test "$use_capabilities" = "yes" ; then
use_capabilities=no
AC_CHECK_HEADERS(sys/capability.h)
if test "$ac_cv_header_sys_capability_h" = "yes" ; then
AC_CHECK_LIB(cap, cap_init, ac_need_libcap=1)
if test "$ac_cv_lib_cap_cap_init" = "yes"; then
AC_DEFINE(USE_CAPABILITIES,1,
[define if capabilities should be used])
LIBS="$LIBS -lcap"
use_capabilities=yes
fi
fi
if test "$use_capabilities" = "no" ; then
AC_MSG_WARN([[
***
*** The use of capabilities on this system is not possible.
*** You need a recent Linux kernel and some patches:
*** fcaps-2.2.9-990610.patch (kernel patch for 2.2.9)
*** fcap-module-990613.tar.gz (kernel module)
*** libcap-1.92.tar.gz (user mode library and utilities)
*** And you have to configure the kernel with CONFIG_VFS_CAP_PLUGIN
*** set (filesystems menu). Be warned: This code is *really* ALPHA.
***]])
fi
fi
GNUPG_CHECK_IPC
if test "$ac_cv_header_sys_shm_h" = "yes"; then
AC_DEFINE(USE_SHM_COPROCESSING,1,
[define if the shared memory interface should be made available])
fi
dnl
dnl check whether we have a random device
dnl
if test "$try_dev_random" = yes ; then
AC_CACHE_CHECK(for random device, ac_cv_have_dev_random,
[if test -r "$NAME_OF_DEV_RANDOM" && test -r "$NAME_OF_DEV_URANDOM" ; then
ac_cv_have_dev_random=yes; else ac_cv_have_dev_random=no; fi])
if test "$ac_cv_have_dev_random" = yes; then
AC_DEFINE(HAVE_DEV_RANDOM,1,
[defined if the system supports a random device] )
fi
else
AC_MSG_CHECKING(for random device)
ac_cv_have_dev_random=no
AC_MSG_RESULT(has been disabled)
fi
dnl
dnl and whether this device supports ioctl
dnl (Note, that we should do a real test here)
dnl
if test "$ac_cv_have_dev_random" = yes ; then
AC_CHECK_HEADERS(linux/random.h)
AC_CACHE_CHECK(for random device ioctl, ac_cv_have_dev_random_ioctl,
[ if test "$ac_cv_header_linux_random_h" = yes ; then
ac_cv_have_dev_random_ioctl=yes;
else
ac_cv_have_dev_random_ioctl=no;
fi
])
if test "$ac_cv_have_dev_random_ioctl" = yes; then
AC_DEFINE(HAVE_DEV_RANDOM_IOCTL,1,
[defined if the random device supports some IOCTLs])
fi
fi
dnl
dnl Figure out the default linkage mode for random modules
dnl
print_egd_notice=no
if test "$use_static_rnd" = default; then
if test "$ac_cv_have_dev_random" = yes; then
static_random_module="rndlinux"
else
case "${target}" in
*-*-mingw32*|*-*-cygwin*)
static_random_module="rndw32"
AC_DEFINE(USE_STATIC_RNDW32,1,
[At some point in the code we need to know that
we use the Windows random module.])
;;
i?86-emx-os2|i?86-*-os2*emx)
static_random_module="rndos2"
;;
m68k-atari-mint)
static_random_module="rndatari"
;;
i?86-*-msdosdjgpp*)
:
;;
*)
static_random_module="rndunix"
print_egd_notice=yes
;;
esac
fi
else
if test "$use_static_rnd" = none; then
:
else
static_random_module="rnd$use_static_rnd"
if test "$use_static_rnd" = "unix"; then
print_egd_notice=yes
fi
fi
fi
if test "$print_egd_notice" = "yes"; then
AC_MSG_WARN([[
***
*** The performance of the UNIX random gatherer module is not very good
*** and it does not keep the entropy pool over multiple invocations of
*** GnuPG. The suggested way to overcome this problem is to use the
***
*** Entropy Gathering Daemon (EGD)
***
*** which provides a entropy source for the whole system. It is written
*** in Perl and available at the GnuPG FTP servers. For more information
*** consult the GnuPG webpages:
***
*** http://www.gnupg.org/download.html#egd
***
*** You may want to run ./configure with --enable-static-rnd=none to be
*** able to load the EGD client code dynamically; this way you can choose
*** between RNDUNIX and RNDEGD without recompilation.
***]])
fi
dnl
dnl Parse the modules list and build the list
dnl of static and dymically linked modules
dnl
dnl (We always need a static rmd160)
static_modules="$static_modules rmd160 $static_random_module"
STATIC_CIPHER_NAMES=""
STATIC_CIPHER_OBJS=""
DYNAMIC_CIPHER_MODS=""
show_dynlink=""
for name in $MODULES_IN_CIPHER; do
x="yes"
for i in $disallowed_modules; do
if test "$name" = "$i" ; then x="no" ; fi
done;
if test $x = yes; then
x="no"
for i in $static_modules; do
if test "$name" = "$i" ; then
x="yes"
fi
done;
if test $x = yes; then
STATIC_CIPHER_NAMES="$STATIC_CIPHER_NAMES $name"
STATIC_CIPHER_OBJS="$STATIC_CIPHER_OBJS $name.o"
else
DYNAMIC_CIPHER_MODS="$DYNAMIC_CIPHER_MODS $name"
show_dynlink="$show_dynlink $name"
fi
fi
done
AC_SUBST(STATIC_CIPHER_OBJS)
AC_SUBST(STATIC_CIPHER_NAMES)
AC_SUBST(DYNAMIC_CIPHER_MODS)
dnl
dnl And build the constructor file
dnl
test -d cipher || mkdir cipher
cat <<G10EOF >cipher/construct.c
/* automatically generated by configure - do not edit */
G10EOF
show_statlink=""
for name in $STATIC_CIPHER_NAMES; do
echo "void ${name}_constructor(void);" >>cipher/construct.c
show_statlink="$show_statlink $name"
done
cat <<G10EOF >>cipher/construct.c
void
cipher_modules_constructor(void)
{
static int done = 0;
if( done )
return;
done = 1;
G10EOF
for name in $STATIC_CIPHER_NAMES; do
echo " ${name}_constructor();" >>cipher/construct.c
done
echo '}' >>cipher/construct.c
dnl
dnl Figure how to link the cipher modules
dnl
AC_SUBST(STATIC_CIPHER_OBJS)
AC_SUBST(DYNAMIC_CIPHER_MODS)
dnl setup assembler stuff
AC_MSG_CHECKING(for mpi assembler functions)
if test -f $srcdir/mpi/config.links ; then
. $srcdir/mpi/config.links
AC_CONFIG_LINKS("$mpi_ln_list")
ac_cv_mpi_extra_asm_modules="$mpi_extra_modules"
ac_cv_mpi_sflags="$mpi_sflags"
ac_cv_mpi_config_done="yes"
AC_MSG_RESULT(done)
else
AC_MSG_RESULT(failed)
AC_MSG_ERROR([mpi/config.links missing!])
fi
MPI_EXTRA_ASM_OBJS=""
show_extraasm=""
if test "$ac_cv_mpi_extra_asm_modules" != ""; then
for i in $ac_cv_mpi_extra_asm_modules; do
show_extraasm="$show_extraasm $i"
MPI_EXTRA_ASM_OBJS="$MPI_EXTRA_ASM_OBJS $i.o"
done
fi
AC_SUBST(MPI_EXTRA_ASM_OBJS)
MPI_SFLAGS="$ac_cv_mpi_sflags"
AC_SUBST(MPI_SFLAGS)
dnl Do we have zlib? Must do it here because Solaris failed
dnl when compiling a conftest (due to the "-lz" from LIBS).
use_local_zlib=yes
if test "$g10_force_zlib" = "yes"; then
:
else
_cppflags="${CPPFLAGS}"
_ldflags="${LDFLAGS}"
AC_ARG_WITH(zlib,
[ --with-zlib=DIR use libz in DIR],[
if test -d "$withval"; then
CPPFLAGS="${CPPFLAGS} -I$withval/include"
LDFLAGS="${LDFLAGS} -L$withval/lib"
fi
])
AC_CHECK_HEADER(zlib.h,
AC_CHECK_LIB(z, deflateInit2_,
use_local_zlib=no
LIBS="$LIBS -lz",
CPPFLAGS=${_cppflags} LDFLAGS=${_ldflags}),
CPPFLAGS=${_cppflags} LDFLAGS=${_ldflags})
fi
if test "$use_local_zlib" = yes ; then
AM_CONDITIONAL(ENABLE_LOCAL_ZLIB, true)
AC_CONFIG_LINKS(zlib.h:zlib/zlib.h zconf.h:zlib/zconf.h )
ZLIBS="../zlib/libzlib.a"
else
AM_CONDITIONAL(ENABLE_LOCAL_ZLIB, false)
ZLIBS=
fi
AC_SUBST(ZLIBS)
# Allow users to append something to the version string without
# flagging it as development version. The user version parts is
# considered everything after a dash.
changequote(,)dnl
tmp_pat='[a-zA-Z]'
changequote([,])dnl
if echo "$VERSION" | sed 's/-.*//' | grep "$tmp_pat" >/dev/null ; then
AC_DEFINE(IS_DEVELOPMENT_VERSION,1,
[Defined if this is not a regular release])
fi
AM_CONDITIONAL(CROSS_COMPILING, test x$cross_compiling = xyes)
GNUPG_CHECK_GNUMAKE
# add some extra libs here so that previous tests don't fail for
# mysterious reasons - the final link step shoudl bail out.
case "${target}" in
*-*-mingw32*)
LIBS="$LIBS -lws2_32"
;;
*)
;;
esac
if test "$GCC" = yes; then
if test "$USE_MAINTAINER_MODE" = "yes"; then
CFLAGS="$CFLAGS -Wall -Wcast-align -Wshadow -Wstrict-prototypes"
else
CFLAGS="$CFLAGS -Wall"
fi
fi
AC_SUBST(NETLIBS)
# Note the \\\\ for backslashes. Autoconf eats one layer, leaving \\
AC_CONFIG_COMMANDS(g10defs.h,[[
cat >g10defs.tmp <<G10EOF
/* Generated automatically by configure */
#ifdef HAVE_DRIVE_LETTERS
#define G10_LOCALEDIR "c:/lib/gnupg/locale"
#define GNUPG_LIBDIR "c:/lib/gnupg"
#define GNUPG_DATADIR "c:/lib/gnupg"
#define GNUPG_HOMEDIR "c:/gnupg"
#else
#define G10_LOCALEDIR "${datadir}/locale"
#define GNUPG_LIBDIR "${libdir}/gnupg"
#define GNUPG_DATADIR "${datadir}/gnupg"
#ifdef __VMS
#define GNUPG_HOMEDIR "/SYS\$LOGIN/gnupg"
#else
#define GNUPG_HOMEDIR "~/.gnupg"
#endif
#endif
/* those are here to be redefined by handcrafted g10defs.h.
Please note that the string version must not contain more
than one character because the using code assumes strlen()==1 */
#ifdef HAVE_DOSISH_SYSTEM
#define DIRSEP_C '\\\\'
#define EXTSEP_C '.'
#define DIRSEP_S "\\\\"
#define EXTSEP_S "."
#else
#define DIRSEP_C '/'
#define EXTSEP_C '.'
#define DIRSEP_S "/"
#define EXTSEP_S "."
#endif
G10EOF
cat mpi/mpi-asm-defs.h >>g10defs.tmp
if cmp -s g10defs.h g10defs.tmp 2>/dev/null; then
echo "g10defs.h is unchanged"
rm -f g10defs.tmp
else
rm -f g10defs.h
mv g10defs.tmp g10defs.h
echo "g10defs.h created"
fi
]],[[
prefix=$prefix
exec_prefix=$exec_prefix
libdir=$libdir
datadir=$datadir
DATADIRNAME=$DATADIRNAME
]])
AC_CONFIG_FILES([
Makefile
intl/Makefile
po/Makefile.in
util/Makefile
mpi/Makefile
cipher/Makefile
g10/Makefile
keyserver/Makefile
keyserver/gpgkeys_mailto
keyserver/gpgkeys_test
doc/Makefile
tools/Makefile
zlib/Makefile
checks/Makefile
])
AC_OUTPUT
# Give some feedback
echo
echo " Configured for: $PRINTABLE_OS_NAME ($target)"
if test -n "$show_dynlink"; then
echo " Dynamically linked modules:$show_dynlink"
fi
if test -n "$show_statlink"; then
echo " Statically linked modules:$show_statlink"
fi
if test -n "$show_extraasm"; then
echo " Extra cpu specific functions:$show_extraasm"
fi
echo

98
gnupg.txt Normal file
View File

@ -0,0 +1,98 @@
%%comments:
Copyright (C) 2000, 2001 Free Software Foundation, Inc.
Permission is granted to copy, distribute and/or modify this document
under the terms of the GNU Free Documentation License, Version 1.1 or
any later version published by the Free Software Foundation; with no
Invariant Sections, with no Front-Cover Texts, and with no Back-Cover
Texts. A copy of the license is included in the file COPYING.
%%name: GnuPG
%%short-description: Complete implementation of the OpenPGP Internet standard.
%%full-description: A complete and free replacement for PGP. The
program does not use any patented algorithms, and can be used as a
filter program. Can handle all OpenPGP messages and messages generated
by PGP 5.0 and newer unless they use the IDEA algorithm.
Supports ElGamal (signature and encrytion), DSA, AES, 3DES, Blowfish,
Twofish, CAST5, MD5, SHA-1, RIPE-MD-160 and TIGER, and has language
support for sixteen different languages.
It is believed to be fully OpenPGP (as defined in RFC2440) conform,
and has integrated support for HKP keyservers. There are a lot of
useful extra features like anonymous message recipients.
%%category: security, administration, email
%%license: GPL
%%maintainer: Werner Koch <wk@gnupg.org>
%%interface: Command line
%%updated: 1 Jun 2001
%%keywords: PGP, security, decryption, encryption, digital signatures
%%programs:
%%GNU: yes
%%web-page: http://www.gnupg.org/
%%support: Paid extension/consulting from http://www.g10code.com
%%doc: English user manual available from:
http://www.gnupg.org/gph/index.html; Spanish user manual available from:
http://www.gnupg.org/gph/index.html; English HOWTO available from
http://www.dewinter.com/gnupg_howto/english/; French HOWTO available
from http://www.gnupg.org/howtos/fr/; German HOWTO available from
http://www.gnupg.org/howtos/de/; Spanish HOWTO available from
http://www.dewinter.com/gnupg_howto/spanish.
%%developers: Matthew Skala, Michael Roth, Niklas Hernaeus, Remi
Guyomarch, Werner Koch <wk@gnupg.org>.
%%contributors: J. Michael Ashley
%%sponsors: German Federal Ministry of Economics and Technology
%%source: ftp://ftp.gnupg.org/gcrypt/gnupg
%%debian: http://www.debian.org/Packages/stable/non-us/gnupg.html
%%redhat: http://www.megaloman.com/~hany/RPM/gnupg.html
%%repository: See http://www.gnupg.org/cvs-access.html
%%related: shred
%%source-language: C
%%supported-languages:
%%use-requirements:
%%build-prerequisites:
%%weak-prerequisites:
%%source-prerequisites:
%%version: 1.0.6 stable released on 29 May 2001
%%announce-list: announce@gnupg.org
%%announce-news:
%%help-list: gnupg-users@gnupg.org
%%help-news:
%%dev-list: gnupg-devel@gnupg.org
%%dev-news:
%%bug-list: bug-gnupg@gnupg.org

325
mpi/ChangeLog Normal file
View File

@ -0,0 +1,325 @@
2002-05-10 Stefan Bellon <sbellon@sbellon.de>
* mpiutil.c (mpi_alloc_like/mpi_debug_alloc_like): Added code
for M_DEBUG.
2002-04-18 Werner Koch <wk@gnupg.org>
* i386/syntax.h (ALIGN): Removed parens from definition
* i386/mpih-add1.S, i386/mpih-sub1.S, i386/mpih-rshift.S,
i386/mpih-lshift.S: Minor syntax changes suggested by Mark Pettit
after comparing the files with those for GMP 4.
2001-11-08 Werner Koch <wk@gnupg.org>
* config.links (mpi_sflags): Add extra rule for sparc64-sun-solaris2.
2001-08-20 Werner Koch <wk@gnupg.org>
* longlong.h [__riscos__]: Need a special pragma here.
2001-08-09 Werner Koch <wk@gnupg.org>
* config.links: Added configuraton for powerpc-openbsd. By Peter
Valchev
2001-07-09 Werner Koch <wk@gnupg.org>
* config.links: Changed the way the list of files to be
symlinked is returned.
2001-05-27 Werner Koch <wk@gnupg.org>
* hppa/, hppa1.1/, pa7100/ : Use .label command instead of labels
because there syntax changed. By Matthew Wilcox.
2001-05-06 Werner Koch <wk@gnupg.org>
* longlong.h: Fixes for ARM by Phil Blundell.
2001-04-17 Werner Koch <wk@gnupg.org>
Updated copyright notices.
2001-03-24 Werner Koch <wk@gnupg.org>
* mpi-mul.c (mpi_mul): Make sure that secret temporary results are
not stored in w. Suggested by Florian Weimer.
2001-03-18 Werner Koch <wk@gnupg.org>
* config.links: Use i386 code for i386. According to tests by
Kevin Ryde the i586 code runs slow on i386 CPUs. Ditto for i786.
2000-10-24 Werner Koch <wk@gnupg.org>
* mips3/: Changed a few comments to C-style. By Jeff Long.
2000-10-13 Werner Koch <wk@gnupg.org>
* mpi.h: Removed the inclusion of mpi-asm-defs.h because this
makes some trouble when doing a VPATH build. configure now
takes care of it.
2000-10-12 Werner Koch <wk@gnupg.org>
* generic/mpi-asm-defs.h: New.
* mips3/mpi-asm-defs.h: New.
* config.links: Create a link to one of the above files.
Wed Jul 19 11:26:43 CEST 2000 Werner Koch <wk@>
* config.links: Support for powerpc--netbsd by Gabriel Rosenkoetter.
Wed Mar 22 13:50:24 CET 2000 Werner Koch <wk@openit.de>
* config.links: Add support for FreeBSD 5 and made the case stmt
looking nicer. From Jun Kuriyama.
Fri Mar 17 17:50:25 CET 2000 Werner Koch <wk@openit.de>
* config.links (sparc64-unknown-linux-gnu): use udic module.
From Adam Mitchell.
2000-03-14 12:03:56 Werner Koch (wk@habibti.openit.de)
* Makefile.am: Do not use .s and .S files but a temp names, so that
OSes with caseinsensitive filenames do work. From Frank Donahoe.
Tue Mar 7 18:45:31 CET 2000 Werner Koch <wk@gnupg.de>
* mpih-mul.c (mpihelp_mul_karatsuba_case): It seems that the
untested part works fine. Removed the debugging message.
* longlong.h (umul_ppmm): Fixes for ARM-4. By Sean MacLennan.
* config.links: Add support for NetBSD.
Thu Jan 13 19:31:58 CET 2000 Werner Koch <wk@gnupg.de>
* mpi-internal.h (karatsuba_ctx): New.
* mpih-mul.c (mpihelp_release_karatsuba_ctx): New.
(mpihelp_mul_karatsuba_case): New.
(mpihelp_mul): Splitted to make use of the new functions.
* mpi-pow.c (mpi_powm): Make use of the new splitted function
to avoid multiple allocation of temporary memory during the
karatsuba operations.
* mpi_mpow.c: Removed the unused Barrett code.
Sun Dec 19 15:22:26 CET 1999 Werner Koch <wk@gnupg.de>
* power/ : Converted more comments to C comments because some AS
complain about ' in comments.
Thu Dec 16 10:07:58 CET 1999 Werner Koch <wk@gnupg.de>
* Makefile.am: c/SFLAGS/ASFLAGS/. This has only been used by the
powerpc and actually never passed the -Wa,foo to the cc.
Thu Dec 9 10:31:05 CET 1999 Werner Koch <wk@gnupg.de>
* power/: Add all files from GMP for this CPU.
* config.links: Support for BSDI 4.x. By Wayne Chapeskie.
(sparc8): Made the search path the same as sparc9
* mpih-div.c (mpihelp_divrem): The MPN_COPY_DECR copied one
elemnat too many. This is gmp2.0.2p9.txt patch.
Sat Oct 9 20:34:41 CEST 1999 Werner Koch <wk@gnupg.de>
* Makefile.am: Removed libtool.
Mon Aug 30 20:38:33 CEST 1999 Werner Koch <wk@isil.d.shuttle.de>
* config.links: Add case label for DJGPP
Wed Jul 14 19:42:08 CEST 1999 Werner Koch <wk@isil.d.shuttle.de>
* Makefile.am: Use .s files as temporaries, disabled other .S rules.
Wed Jul 7 13:08:40 CEST 1999 Werner Koch <wk@isil.d.shuttle.de>
* mpicoder.c (g10_log_mpidump): New.
* Makefile.am: Support for libtool.
Fri Jul 2 11:45:54 CEST 1999 Werner Koch <wk@isil.d.shuttle.de>
* mpi-bit.c (mpi_lshift_limbs,mpi_rshift_limbs): New.
* mpi-mpow.c (barrett_mulm): New but diabled.
Tue Jun 1 16:01:46 CEST 1999 Werner Koch <wk@isil.d.shuttle.de>
* config.links (i[56]86*-*-freebsdelf*): New.
Sun May 23 14:20:22 CEST 1999 Werner Koch <wk@isil.d.shuttle.de>
* config.links (sysdep.h): Not any more conditionally created.
Tue May 4 15:47:53 CEST 1999 Werner Koch <wk@isil.d.shuttle.de>
* mpiutil.c (mpi_alloc_like): New.
Mon Apr 26 17:48:15 CEST 1999 Werner Koch <wk@isil.d.shuttle.de>
* mpih-add.c, mpih-sub.c: Removed
* mpi-inline.c: New.
* mpi-inline.h: Make it usable by mpi-inline.c.
Sun Apr 18 10:11:28 CEST 1999 Werner Koch <wk@isil.d.shuttle.de>
* mpih-mul.c (mpihelp_mul_n): Fixed use of memory region.
(mpihelp_mul): Ditto.
Wed Apr 7 20:51:39 CEST 1999 Werner Koch <wk@isil.d.shuttle.de>
* Makefile.am: Explicit rules to invoke cpp on *.S
Mon Mar 8 20:47:17 CET 1999 Werner Koch <wk@isil.d.shuttle.de>
* config.links: Take advantage of the with_symbol_underscore macro.
Add support for freebsd 4.
Wed Feb 24 11:07:27 CET 1999 Werner Koch <wk@isil.d.shuttle.de>
* mips3/mpih-sub1.S: Removed left over junk in last line. (Should I
blame me or my editor?).
Sat Feb 13 12:04:43 CET 1999 Werner Koch <wk@isil.d.shuttle.de>
* Makefile.am: Removed the +=. Add MPI_OPT_FLAGS.
Sat Jan 9 16:02:23 CET 1999 Werner Koch <wk@isil.d.shuttle.de>
* mpi-cmp.c (mpi_cmp_ui): Normalized the arg.
Thu Jan 7 18:00:58 CET 1999 Werner Koch <wk@isil.d.shuttle.de>
* mpi-bit.c (mpi_normalize): New.
(mpi_get_nbits): Normalize the MPI.
* mpi-bit.c (mpi_cmp): Normalize the MPI before the compare.
Tue Dec 8 13:15:16 CET 1998 Werner Koch <wk@isil.d.shuttle.de>
* config.links: Moved the case for powerpc*linux
* powerpcp32/*.S: Removed some underscores.
Thu Nov 26 07:27:52 1998 Werner Koch <werner.koch@guug.de>
* config.links: Support for ppc with ELF
* powerpc32/syntax.h: New.
* powerpc32/*.S: Applied ELF patches (glibc patches)
Tue Nov 10 19:31:37 1998 Werner Koch (wk@isil.d.shuttle.de)
* power*/ : Started with stuff for PPC
* config.links: Some stuff for PPC.
* generic/udiv-w-sdiv.c: New but disabled.
Tue Oct 27 12:37:46 1998 Werner Koch (wk@isil.d.shuttle.de)
* config.links (freebsd): Fixes for FreeBSD 3.0
Wed Oct 14 09:59:30 1998 Werner Koch (wk@isil.d.shuttle.de)
* config.links (freebsd): ELF patches from Jun Kuriyama.
Thu Oct 8 13:28:17 1998 Werner Koch (wk@isil.d.shuttle.de)
* mpi-mpow.c (mpi_mulpowm): Fixed mem leak (m_free/mpi_free).
Thu Sep 17 18:08:50 1998 Werner Koch (wk@(none))
* hppa1.1/udiv-qrnnd.S: Fix from Steffen Zahn for HPUX 10.20
Thu Aug 6 16:39:28 1998 Werner Koch,mobil,,, (wk@tobold)
* mpi-bit.c (mpi_set_bytes): Removed.
Wed Aug 5 15:11:12 1998 Werner Koch (wk@(none))
* mpicoder.c (mpi_read_from_buffer): New.
* mpiutil.c (mpi_set_opaque): New.
(mpi_get_opaque): New.
(mpi_copy): Changed to support opauqe flag
(mpi_free): Ditto.
Sat Jul 4 10:11:11 1998 Werner Koch (wk@isil.d.shuttle.de)
* mpiutil.c (mpi_clear): Reset flags.
(mpi_set): Ditto.
(mpi_alloc_secure): Set flag to 1 and not ored the 1 in, tsss..
Fri Jun 26 11:19:06 1998 Werner Koch (wk@isil.d.shuttle.de)
* mpiutil.c (mpi_alloc): set nbits to 0.
(mpi_alloc_secure): Ditto.
(mpi_clear): Ditto.
Thu Jun 25 11:50:01 1998 Werner Koch (wk@isil.d.shuttle.de)
* mips3/*.S: New
Mon May 18 13:47:06 1998 Werner Koch (wk@isil.d.shuttle.de)
* config.links: split mpih-shift into mpih-[lr]shift and
changed all implementations.
* mpi/alpha: add some new assembler stuff.
Wed May 13 11:04:29 1998 Werner Koch (wk@isil.d.shuttle.de)
* config.links: Add support for MIPS
Thu Apr 9 11:31:36 1998 Werner Koch (wk@isil.d.shuttle.de)
* mpicoder.c (mpi_get_secure_buffer): New.
Wed Apr 8 09:44:33 1998 Werner Koch (wk@isil.d.shuttle.de)
* config.links: Applied small fix from Ulf Möller.
Mon Apr 6 12:38:52 1998 Werner Koch (wk@isil.d.shuttle.de)
* mpicoder.c (mpi_get_buffer): Removed returned leading zeroes
and changed all callers.
Tue Mar 10 13:40:34 1998 Werner Koch (wk@isil.d.shuttle.de)
* mpi-bit.c (mpi_clear_highbit): New.
Mon Mar 2 19:29:00 1998 Werner Koch (wk@isil.d.shuttle.de)
* Makefile.am (DISTCLEANFILES): New
Thu Feb 26 06:48:54 1998 Werner Koch (wk@isil.d.shuttle.de)
* config.links (X86_BROKEN_ALIGN): Added for some systems.
Mon Feb 23 12:21:40 1998 Werner Koch (wk@isil.d.shuttle.de)
* mpi/m68k/mpih-shift.S (Lspecial): Changed duplicate symbol.
Mon Feb 16 13:00:27 1998 Werner Koch (wk@isil.d.shuttle.de)
* config.links : Add detection of m68k cpus
Copyright 1998, 1999, 2000, 2001 Free Software Foundation, Inc.
This file is free software; as a special exception the author gives
unlimited permission to copy and/or distribute it, with or without
modifications, as long as this notice is preserved.
This file is distributed in the hope that it will be useful, but
WITHOUT ANY WARRANTY, to the extent permitted by law; without even the
implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.

79
mpi/Makefile.am Normal file
View File

@ -0,0 +1,79 @@
# Copyright (C) 1998, 1999, 2000, 2001 Free Software Foundation, Inc.
#
# This file is part of GnuPG.
#
# GnuPG is free software; you can redistribute it and/or modify
# it under the terms of the GNU General Public License as published by
# the Free Software Foundation; either version 2 of the License, or
# (at your option) any later version.
#
# GnuPG is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with this program; if not, write to the Free Software
# Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA
## Process this file with automake to produce Makefile.in
INCLUDES = -I.. -I$(top_srcdir)/include
CFLAGS = @CFLAGS@ @MPI_OPT_FLAGS@
ASFLAGS = @MPI_SFLAGS@
EXTRA_DIST = config.links
DISTCLEANFILES = mpi-asm-defs.h \
mpih-add1.S mpih-mul1.S mpih-mul2.S mpih-mul3.S \
mpih-lshift.S mpih-rshift.S mpih-sub1.S asm-syntax.h sysdep.h
# Note: we only use .S files so we should delete all left over .s
CLEANFILES = _*.s
noinst_LIBRARIES = libmpi.a
# libmpi_a_LDFLAGS =
libmpi_a_SOURCES = longlong.h \
mpi-add.c \
mpi-bit.c \
mpi-cmp.c \
mpi-div.c \
mpi-gcd.c \
mpi-internal.h \
mpi-inline.h \
mpi-inline.c \
mpi-inv.c \
mpi-mul.c \
mpi-pow.c \
mpi-mpow.c \
mpi-scan.c \
mpicoder.c \
mpih-cmp.c \
mpih-div.c \
mpih-mul.c \
mpiutil.c \
g10m.c
# Note this objects are actually links, the sourcefiles are
# distributed by special code in dist-hook
common_asm_objects = mpih-mul1.o \
mpih-mul2.o \
mpih-mul3.o \
mpih-add1.o \
mpih-sub1.o \
mpih-lshift.o \
mpih-rshift.o
libmpi_a_DEPENDENCIES = $(common_asm_objects) @MPI_EXTRA_ASM_OBJS@
libmpi_a_LIBADD = $(common_asm_objects) @MPI_EXTRA_ASM_OBJS@
# cancel the default rules used by libtool which do not really
# work and add one to cpp .S files
.S.o:
$(CPP) $(INCLUDES) $(DEFS) $< | grep -v '^#' > _$*.s
$(COMPILE) -c _$*.s
mv -f _$*.o $*.o
.S.lo:

53
mpi/alpha/README Normal file
View File

@ -0,0 +1,53 @@
This directory contains mpn functions optimized for DEC Alpha processors.
RELEVANT OPTIMIZATION ISSUES
EV4
1. This chip has very limited store bandwidth. The on-chip L1 cache is
write-through, and a cache line is transfered from the store buffer to the
off-chip L2 in as much 15 cycles on most systems. This delay hurts
mpn_add_n, mpn_sub_n, mpn_lshift, and mpn_rshift.
2. Pairing is possible between memory instructions and integer arithmetic
instructions.
3. mulq and umulh is documented to have a latency of 23 cycles, but 2 of
these cycles are pipelined. Thus, multiply instructions can be issued at a
rate of one each 21nd cycle.
EV5
1. The memory bandwidth of this chip seems excellent, both for loads and
stores. Even when the working set is larger than the on-chip L1 and L2
caches, the perfromance remain almost unaffected.
2. mulq has a measured latency of 13 cycles and an issue rate of 1 each 8th
cycle. umulh has a measured latency of 15 cycles and an issue rate of 1
each 10th cycle. But the exact timing is somewhat confusing.
3. mpn_add_n. With 4-fold unrolling, we need 37 instructions, whereof 12
are memory operations. This will take at least
ceil(37/2) [dual issue] + 1 [taken branch] = 20 cycles
We have 12 memory cycles, plus 4 after-store conflict cycles, or 16 data
cache cycles, which should be completely hidden in the 20 issue cycles.
The computation is inherently serial, with these dependencies:
addq
/ \
addq cmpult
| |
cmpult |
\ /
or
I.e., there is a 4 cycle path for each limb, making 16 cycles the absolute
minimum. We could replace the `or' with a cmoveq/cmovne, which would save
a cycle on EV5, but that might waste a cycle on EV4. Also, cmov takes 2
cycles.
addq
/ \
addq cmpult
| \
cmpult -> cmovne
STATUS

11
mpi/alpha/distfiles Normal file
View File

@ -0,0 +1,11 @@
README
mpih-add1.S
mpih-sub1.S
mpih-mul1.S
mpih-mul2.S
mpih-mul3.S
mpih-lshift.S
mpih-rshift.S
udiv-qrnnd.S

124
mpi/alpha/mpih-add1.S Normal file
View File

@ -0,0 +1,124 @@
/* alpha add_n -- Add two limb vectors of the same length > 0 and store
* sum in a third limb vector.
* Copyright (C) 1995, 1998, 2001 Free Software Foundation, Inc.
*
* This file is part of GnuPG.
*
* GnuPG is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* GnuPG is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA
*/
/*******************
* mpi_limb_t
* mpihelp_add_n( mpi_ptr_t res_ptr, ($16)
* mpi_ptr_t s1_ptr, ($17)
* mpi_ptr_t s2_ptr, ($18)
* mpi_size_t size) ($19)
*/
.set noreorder
.set noat
.text
.align 3
.globl mpihelp_add_n
.ent mpihelp_add_n
mpihelp_add_n:
.frame $30,0,$26,0
ldq $3,0($17)
ldq $4,0($18)
subq $19,1,$19
and $19,4-1,$2 # number of limbs in first loop
bis $31,$31,$0
beq $2,.L0 # if multiple of 4 limbs, skip first loop
subq $19,$2,$19
.Loop0: subq $2,1,$2
ldq $5,8($17)
addq $4,$0,$4
ldq $6,8($18)
cmpult $4,$0,$1
addq $3,$4,$4
cmpult $4,$3,$0
stq $4,0($16)
or $0,$1,$0
addq $17,8,$17
addq $18,8,$18
bis $5,$5,$3
bis $6,$6,$4
addq $16,8,$16
bne $2,.Loop0
.L0: beq $19,.Lend
.align 3
.Loop: subq $19,4,$19
ldq $5,8($17)
addq $4,$0,$4
ldq $6,8($18)
cmpult $4,$0,$1
addq $3,$4,$4
cmpult $4,$3,$0
stq $4,0($16)
or $0,$1,$0
ldq $3,16($17)
addq $6,$0,$6
ldq $4,16($18)
cmpult $6,$0,$1
addq $5,$6,$6
cmpult $6,$5,$0
stq $6,8($16)
or $0,$1,$0
ldq $5,24($17)
addq $4,$0,$4
ldq $6,24($18)
cmpult $4,$0,$1
addq $3,$4,$4
cmpult $4,$3,$0
stq $4,16($16)
or $0,$1,$0
ldq $3,32($17)
addq $6,$0,$6
ldq $4,32($18)
cmpult $6,$0,$1
addq $5,$6,$6
cmpult $6,$5,$0
stq $6,24($16)
or $0,$1,$0
addq $17,32,$17
addq $18,32,$18
addq $16,32,$16
bne $19,.Loop
.Lend: addq $4,$0,$4
cmpult $4,$0,$1
addq $3,$4,$4
cmpult $4,$3,$0
stq $4,0($16)
or $0,$1,$0
ret $31,($26),1
.end mpihelp_add_n

123
mpi/alpha/mpih-lshift.S Normal file
View File

@ -0,0 +1,123 @@
/* alpha - left shift
*
* Copyright (C) 1994, 1995, 1998, 2001 Free Software Foundation, Inc.
*
*
* This file is part of GnuPG.
*
* GnuPG is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* GnuPG is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA
*
* Note: This code is heavily based on the GNU MP Library.
* Actually it's the same code with only minor changes in the
* way the data is stored; this is to support the abstraction
* of an optional secure memory allocation which may be used
* to avoid revealing of sensitive data due to paging etc.
* The GNU MP Library itself is published under the LGPL;
* however I decided to publish this code under the plain GPL.
*/
/*******************
* mpi_limb_t
* mpihelp_lshift( mpi_ptr_t wp, (r16)
* mpi_ptr_t up, (r17)
* mpi_size_t usize, (r18)
* unsigned cnt) (r19)
*
* This code runs at 4.8 cycles/limb on the 21064. With infinite unrolling,
* it would take 4 cycles/limb. It should be possible to get down to 3
* cycles/limb since both ldq and stq can be paired with the other used
* instructions. But there are many restrictions in the 21064 pipeline that
* makes it hard, if not impossible, to get down to 3 cycles/limb:
*
* 1. ldq has a 3 cycle delay, srl and sll have a 2 cycle delay.
* 2. Only aligned instruction pairs can be paired.
* 3. The store buffer or silo might not be able to deal with the bandwidth.
*/
.set noreorder
.set noat
.text
.align 3
.globl mpihelp_lshift
.ent mpihelp_lshift
mpihelp_lshift:
.frame $30,0,$26,0
s8addq $18,$17,$17 # make r17 point at end of s1
ldq $4,-8($17) # load first limb
subq $17,8,$17
subq $31,$19,$7
s8addq $18,$16,$16 # make r16 point at end of RES
subq $18,1,$18
and $18,4-1,$20 # number of limbs in first loop
srl $4,$7,$0 # compute function result
beq $20,.L0
subq $18,$20,$18
.align 3
.Loop0:
ldq $3,-8($17)
subq $16,8,$16
subq $17,8,$17
subq $20,1,$20
sll $4,$19,$5
srl $3,$7,$6
bis $3,$3,$4
bis $5,$6,$8
stq $8,0($16)
bne $20,.Loop0
.L0: beq $18,.Lend
.align 3
.Loop: ldq $3,-8($17)
subq $16,32,$16
subq $18,4,$18
sll $4,$19,$5
srl $3,$7,$6
ldq $4,-16($17)
sll $3,$19,$1
bis $5,$6,$8
stq $8,24($16)
srl $4,$7,$2
ldq $3,-24($17)
sll $4,$19,$5
bis $1,$2,$8
stq $8,16($16)
srl $3,$7,$6
ldq $4,-32($17)
sll $3,$19,$1
bis $5,$6,$8
stq $8,8($16)
srl $4,$7,$2
subq $17,32,$17
bis $1,$2,$8
stq $8,0($16)
bgt $18,.Loop
.Lend: sll $4,$19,$8
stq $8,-8($16)
ret $31,($26),1
.end mpihelp_lshift

90
mpi/alpha/mpih-mul1.S Normal file
View File

@ -0,0 +1,90 @@
/* Alpha 21064 mpih-mul1.S -- Multiply a limb vector with a limb and store
* the result in a second limb vector.
*
* Copyright (C) 1992, 1994, 1995, 1998,
2001 Free Software Foundation, Inc.
*
* This file is part of GnuPG.
*
* GnuPG is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* GnuPG is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA
*/
/*******************
* mpi_limb_t
* mpihelp_mul_1( mpi_ptr_t res_ptr, (r16)
* mpi_ptr_t s1_ptr, (r17)
* mpi_size_t s1_size, (r18)
* mpi_limb_t s2_limb) (r19)
*
* This code runs at 42 cycles/limb on the EV4 and 18 cycles/limb on the EV5.
*
* To improve performance for long multiplications, we would use
* 'fetch' for S1 and 'fetch_m' for RES. It's not obvious how to use
* these instructions without slowing down the general code: 1. We can
* only have two prefetches in operation at any time in the Alpha
* architecture. 2. There will seldom be any special alignment
* between RES_PTR and S1_PTR. Maybe we can simply divide the current
* loop into an inner and outer loop, having the inner loop handle
* exactly one prefetch block?
*/
.set noreorder
.set noat
.text
.align 3
.globl mpihelp_mul_1
.ent mpihelp_mul_1 2
mpihelp_mul_1:
.frame $30,0,$26
ldq $2,0($17) # $2 = s1_limb
subq $18,1,$18 # size--
mulq $2,$19,$3 # $3 = prod_low
bic $31,$31,$4 # clear cy_limb
umulh $2,$19,$0 # $0 = prod_high
beq $18,Lend1 # jump if size was == 1
ldq $2,8($17) # $2 = s1_limb
subq $18,1,$18 # size--
stq $3,0($16)
beq $18,Lend2 # jump if size was == 2
.align 3
Loop: mulq $2,$19,$3 # $3 = prod_low
addq $4,$0,$0 # cy_limb = cy_limb + 'cy'
subq $18,1,$18 # size--
umulh $2,$19,$4 # $4 = cy_limb
ldq $2,16($17) # $2 = s1_limb
addq $17,8,$17 # s1_ptr++
addq $3,$0,$3 # $3 = cy_limb + prod_low
stq $3,8($16)
cmpult $3,$0,$0 # $0 = carry from (cy_limb + prod_low)
addq $16,8,$16 # res_ptr++
bne $18,Loop
Lend2: mulq $2,$19,$3 # $3 = prod_low
addq $4,$0,$0 # cy_limb = cy_limb + 'cy'
umulh $2,$19,$4 # $4 = cy_limb
addq $3,$0,$3 # $3 = cy_limb + prod_low
cmpult $3,$0,$0 # $0 = carry from (cy_limb + prod_low)
stq $3,8($16)
addq $4,$0,$0 # cy_limb = prod_high + cy
ret $31,($26),1
Lend1: stq $3,0($16)
ret $31,($26),1
.end mpihelp_mul_1

97
mpi/alpha/mpih-mul2.S Normal file
View File

@ -0,0 +1,97 @@
/* Alpha 21064 addmul_1 -- Multiply a limb vector with a limb and add
* the result to a second limb vector.
*
* Copyright (C) 1992, 1994, 1995, 1998,
* 2001 Free Software Foundation, Inc.
*
* This file is part of GnuPG.
*
* GnuPG is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* GnuPG is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA
*/
/*******************
* mpi_limb_t
* mpihelp_addmul_1( mpi_ptr_t res_ptr, (r16)
* mpi_ptr_t s1_ptr, (r17)
* mpi_size_t s1_size, (r18)
* mpi_limb_t s2_limb) (r19)
*
* This code runs at 42 cycles/limb on EV4 and 18 cycles/limb on EV5.
*/
.set noreorder
.set noat
.text
.align 3
.globl mpihelp_addmul_1
.ent mpihelp_addmul_1 2
mpihelp_addmul_1:
.frame $30,0,$26
ldq $2,0($17) # $2 = s1_limb
addq $17,8,$17 # s1_ptr++
subq $18,1,$18 # size--
mulq $2,$19,$3 # $3 = prod_low
ldq $5,0($16) # $5 = *res_ptr
umulh $2,$19,$0 # $0 = prod_high
beq $18,.Lend1 # jump if size was == 1
ldq $2,0($17) # $2 = s1_limb
addq $17,8,$17 # s1_ptr++
subq $18,1,$18 # size--
addq $5,$3,$3
cmpult $3,$5,$4
stq $3,0($16)
addq $16,8,$16 # res_ptr++
beq $18,.Lend2 # jump if size was == 2
.align 3
.Loop: mulq $2,$19,$3 # $3 = prod_low
ldq $5,0($16) # $5 = *res_ptr
addq $4,$0,$0 # cy_limb = cy_limb + 'cy'
subq $18,1,$18 # size--
umulh $2,$19,$4 # $4 = cy_limb
ldq $2,0($17) # $2 = s1_limb
addq $17,8,$17 # s1_ptr++
addq $3,$0,$3 # $3 = cy_limb + prod_low
cmpult $3,$0,$0 # $0 = carry from (cy_limb + prod_low)
addq $5,$3,$3
cmpult $3,$5,$5
stq $3,0($16)
addq $16,8,$16 # res_ptr++
addq $5,$0,$0 # combine carries
bne $18,.Loop
.Lend2: mulq $2,$19,$3 # $3 = prod_low
ldq $5,0($16) # $5 = *res_ptr
addq $4,$0,$0 # cy_limb = cy_limb + 'cy'
umulh $2,$19,$4 # $4 = cy_limb
addq $3,$0,$3 # $3 = cy_limb + prod_low
cmpult $3,$0,$0 # $0 = carry from (cy_limb + prod_low)
addq $5,$3,$3
cmpult $3,$5,$5
stq $3,0($16)
addq $5,$0,$0 # combine carries
addq $4,$0,$0 # cy_limb = prod_high + cy
ret $31,($26),1
.Lend1: addq $5,$3,$3
cmpult $3,$5,$5
stq $3,0($16)
addq $0,$5,$0
ret $31,($26),1
.end mpihelp_addmul_1

95
mpi/alpha/mpih-mul3.S Normal file
View File

@ -0,0 +1,95 @@
/* Alpha 21064 submul_1 -- Multiply a limb vector with a limb and
* subtract the result from a second limb vector.
* Copyright (C) 1992, 1994, 1995, 1998,
* 2001 Free Software Foundation, Inc.
*
* This file is part of GnuPG.
*
* GnuPG is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* GnuPG is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA
*/
/*******************
* mpi_limb_t
* mpihelp_submul_1( mpi_ptr_t res_ptr, (r16 )
* mpi_ptr_t s1_ptr, (r17 )
* mpi_size_t s1_size, (r18 )
* mpi_limb_t s2_limb) (r19 )
*
* This code runs at 42 cycles/limb on EV4 and 18 cycles/limb on EV5.
*/
.set noreorder
.set noat
.text
.align 3
.globl mpihelp_submul_1
.ent mpihelp_submul_1 2
mpihelp_submul_1:
.frame $30,0,$26
ldq $2,0($17) # $2 = s1_limb
addq $17,8,$17 # s1_ptr++
subq $18,1,$18 # size--
mulq $2,$19,$3 # $3 = prod_low
ldq $5,0($16) # $5 = *res_ptr
umulh $2,$19,$0 # $0 = prod_high
beq $18,.Lend1 # jump if size was == 1
ldq $2,0($17) # $2 = s1_limb
addq $17,8,$17 # s1_ptr++
subq $18,1,$18 # size--
subq $5,$3,$3
cmpult $5,$3,$4
stq $3,0($16)
addq $16,8,$16 # res_ptr++
beq $18,.Lend2 # jump if size was == 2
.align 3
.Loop: mulq $2,$19,$3 # $3 = prod_low
ldq $5,0($16) # $5 = *res_ptr
addq $4,$0,$0 # cy_limb = cy_limb + 'cy'
subq $18,1,$18 # size--
umulh $2,$19,$4 # $4 = cy_limb
ldq $2,0($17) # $2 = s1_limb
addq $17,8,$17 # s1_ptr++
addq $3,$0,$3 # $3 = cy_limb + prod_low
cmpult $3,$0,$0 # $0 = carry from (cy_limb + prod_low)
subq $5,$3,$3
cmpult $5,$3,$5
stq $3,0($16)
addq $16,8,$16 # res_ptr++
addq $5,$0,$0 # combine carries
bne $18,.Loop
.Lend2: mulq $2,$19,$3 # $3 = prod_low
ldq $5,0($16) # $5 = *res_ptr
addq $4,$0,$0 # cy_limb = cy_limb + 'cy'
umulh $2,$19,$4 # $4 = cy_limb
addq $3,$0,$3 # $3 = cy_limb + prod_low
cmpult $3,$0,$0 # $0 = carry from (cy_limb + prod_low)
subq $5,$3,$3
cmpult $5,$3,$5
stq $3,0($16)
addq $5,$0,$0 # combine carries
addq $4,$0,$0 # cy_limb = prod_high + cy
ret $31,($26),1
.Lend1: subq $5,$3,$3
cmpult $5,$3,$5
stq $3,0($16)
addq $0,$5,$0
ret $31,($26),1
.end mpihelp_submul_1

120
mpi/alpha/mpih-rshift.S Normal file
View File

@ -0,0 +1,120 @@
/* alpha rshift
* Copyright (C) 1994, 1995, 1998, 1999,
* 2000, 2001 Free Software Foundation, Inc.
*
* This file is part of GNUPG.
*
* GNUPG is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* GNUPG is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA
*
* Note: This code is heavily based on the GNU MP Library.
* Actually it's the same code with only minor changes in the
* way the data is stored; this is to support the abstraction
* of an optional secure memory allocation which may be used
* to avoid revealing of sensitive data due to paging etc.
* The GNU MP Library itself is published under the LGPL;
* however I decided to publish this code under the plain GPL.
*/
/*******************
* mpi_limb_t
* mpihelp_rshift( mpi_ptr_t wp, (r16)
* mpi_ptr_t up, (r17)
* mpi_size_t usize, (r18)
* unsigned cnt) (r19)
*
* This code runs at 4.8 cycles/limb on the 21064. With infinite unrolling,
* it would take 4 cycles/limb. It should be possible to get down to 3
* cycles/limb since both ldq and stq can be paired with the other used
* instructions. But there are many restrictions in the 21064 pipeline that
* makes it hard, if not impossible, to get down to 3 cycles/limb:
*
* 1. ldq has a 3 cycle delay, srl and sll have a 2 cycle delay.
* 2. Only aligned instruction pairs can be paired.
* 3. The store buffer or silo might not be able to deal with the bandwidth.
*/
.set noreorder
.set noat
.text
.align 3
.globl mpihelp_rshift
.ent mpihelp_rshift
mpihelp_rshift:
.frame $30,0,$26,0
ldq $4,0($17) # load first limb
addq $17,8,$17
subq $31,$19,$7
subq $18,1,$18
and $18,4-1,$20 # number of limbs in first loop
sll $4,$7,$0 # compute function result
beq $20,.R0
subq $18,$20,$18
.align 3
.Roop0:
ldq $3,0($17)
addq $16,8,$16
addq $17,8,$17
subq $20,1,$20
srl $4,$19,$5
sll $3,$7,$6
bis $3,$3,$4
bis $5,$6,$8
stq $8,-8($16)
bne $20,.Roop0
.R0: beq $18,.Rend
.align 3
.Roop: ldq $3,0($17)
addq $16,32,$16
subq $18,4,$18
srl $4,$19,$5
sll $3,$7,$6
ldq $4,8($17)
srl $3,$19,$1
bis $5,$6,$8
stq $8,-32($16)
sll $4,$7,$2
ldq $3,16($17)
srl $4,$19,$5
bis $1,$2,$8
stq $8,-24($16)
sll $3,$7,$6
ldq $4,24($17)
srl $3,$19,$1
bis $5,$6,$8
stq $8,-16($16)
sll $4,$7,$2
addq $17,32,$17
bis $1,$2,$8
stq $8,-8($16)
bgt $18,.Roop
.Rend: srl $4,$19,$8
stq $8,0($16)
ret $31,($26),1
.end mpihelp_rshift

124
mpi/alpha/mpih-sub1.S Normal file
View File

@ -0,0 +1,124 @@
/* Alpha sub_n -- Subtract two limb vectors of the same length > 0 and
* store difference in a third limb vector.
* Copyright (C) 1995, 1998,
* 2001 Free Software Foundation, Inc.
*
* This file is part of GnuPG.
*
* GnuPG is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* GnuPG is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA
*/
/*******************
* mpi_limb_t
* mpihelp_sub_n( mpi_ptr_t res_ptr, (r16)
* mpi_ptr_t s1_ptr, (r17)
* mpi_ptr_t s2_ptr, (r18)
* mpi_size_t size) (r19)
*/
.set noreorder
.set noat
.text
.align 3
.globl mpihelp_sub_n
.ent mpihelp_sub_n
mpihelp_sub_n:
.frame $30,0,$26,0
ldq $3,0($17)
ldq $4,0($18)
subq $19,1,$19
and $19,4-1,$2 # number of limbs in first loop
bis $31,$31,$0
beq $2,.L0 # if multiple of 4 limbs, skip first loop
subq $19,$2,$19
.Loop0: subq $2,1,$2
ldq $5,8($17)
addq $4,$0,$4
ldq $6,8($18)
cmpult $4,$0,$1
subq $3,$4,$4
cmpult $3,$4,$0
stq $4,0($16)
or $0,$1,$0
addq $17,8,$17
addq $18,8,$18
bis $5,$5,$3
bis $6,$6,$4
addq $16,8,$16
bne $2,.Loop0
.L0: beq $19,.Lend
.align 3
.Loop: subq $19,4,$19
ldq $5,8($17)
addq $4,$0,$4
ldq $6,8($18)
cmpult $4,$0,$1
subq $3,$4,$4
cmpult $3,$4,$0
stq $4,0($16)
or $0,$1,$0
ldq $3,16($17)
addq $6,$0,$6
ldq $4,16($18)
cmpult $6,$0,$1
subq $5,$6,$6
cmpult $5,$6,$0
stq $6,8($16)
or $0,$1,$0
ldq $5,24($17)
addq $4,$0,$4
ldq $6,24($18)
cmpult $4,$0,$1
subq $3,$4,$4
cmpult $3,$4,$0
stq $4,16($16)
or $0,$1,$0
ldq $3,32($17)
addq $6,$0,$6
ldq $4,32($18)
cmpult $6,$0,$1
subq $5,$6,$6
cmpult $5,$6,$0
stq $6,24($16)
or $0,$1,$0
addq $17,32,$17
addq $18,32,$18
addq $16,32,$16
bne $19,.Loop
.Lend: addq $4,$0,$4
cmpult $4,$0,$1
subq $3,$4,$4
cmpult $3,$4,$0
stq $4,0($16)
or $0,$1,$0
ret $31,($26),1
.end mpihelp_sub_n

161
mpi/alpha/udiv-qrnnd.S Normal file
View File

@ -0,0 +1,161 @@
/* Alpha 21064 __udiv_qrnnd
*
* Copyright (C) 1992, 1994, 1995, 1998,
* 2001 Free Software Foundation, Inc.
*
* This file is part of GnuPG.
*
* GnuPG is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* GnuPG is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA
*
* Note: This code is heavily based on the GNU MP Library.
* Actually it's the same code with only minor changes in the
* way the data is stored; this is to support the abstraction
* of an optional secure memory allocation which may be used
* to avoid revealing of sensitive data due to paging etc.
* The GNU MP Library itself is published under the LGPL;
* however I decided to publish this code under the plain GPL.
*/
.set noreorder
.set noat
.text
.align 3
.globl __udiv_qrnnd
.ent __udiv_qrnnd
__udiv_qrnnd:
.frame $30,0,$26,0
.prologue 0
#define cnt $2
#define tmp $3
#define rem_ptr $16
#define n1 $17
#define n0 $18
#define d $19
#define qb $20
ldiq cnt,16
blt d,.Largedivisor
.Loop1: cmplt n0,0,tmp
addq n1,n1,n1
bis n1,tmp,n1
addq n0,n0,n0
cmpule d,n1,qb
subq n1,d,tmp
cmovne qb,tmp,n1
bis n0,qb,n0
cmplt n0,0,tmp
addq n1,n1,n1
bis n1,tmp,n1
addq n0,n0,n0
cmpule d,n1,qb
subq n1,d,tmp
cmovne qb,tmp,n1
bis n0,qb,n0
cmplt n0,0,tmp
addq n1,n1,n1
bis n1,tmp,n1
addq n0,n0,n0
cmpule d,n1,qb
subq n1,d,tmp
cmovne qb,tmp,n1
bis n0,qb,n0
cmplt n0,0,tmp
addq n1,n1,n1
bis n1,tmp,n1
addq n0,n0,n0
cmpule d,n1,qb
subq n1,d,tmp
cmovne qb,tmp,n1
bis n0,qb,n0
subq cnt,1,cnt
bgt cnt,.Loop1
stq n1,0(rem_ptr)
bis $31,n0,$0
ret $31,($26),1
.Largedivisor:
and n0,1,$4
srl n0,1,n0
sll n1,63,tmp
or tmp,n0,n0
srl n1,1,n1
and d,1,$6
srl d,1,$5
addq $5,$6,$5
.Loop2: cmplt n0,0,tmp
addq n1,n1,n1
bis n1,tmp,n1
addq n0,n0,n0
cmpule $5,n1,qb
subq n1,$5,tmp
cmovne qb,tmp,n1
bis n0,qb,n0
cmplt n0,0,tmp
addq n1,n1,n1
bis n1,tmp,n1
addq n0,n0,n0
cmpule $5,n1,qb
subq n1,$5,tmp
cmovne qb,tmp,n1
bis n0,qb,n0
cmplt n0,0,tmp
addq n1,n1,n1
bis n1,tmp,n1
addq n0,n0,n0
cmpule $5,n1,qb
subq n1,$5,tmp
cmovne qb,tmp,n1
bis n0,qb,n0
cmplt n0,0,tmp
addq n1,n1,n1
bis n1,tmp,n1
addq n0,n0,n0
cmpule $5,n1,qb
subq n1,$5,tmp
cmovne qb,tmp,n1
bis n0,qb,n0
subq cnt,1,cnt
bgt cnt,.Loop2
addq n1,n1,n1
addq $4,n1,n1
bne $6,.LOdd
stq n1,0(rem_ptr)
bis $31,n0,$0
ret $31,($26),1
.LOdd:
/* q' in n0. r' in n1 */
addq n1,n0,n1
cmpult n1,n0,tmp # tmp := carry from addq
beq tmp,.LLp6
addq n0,1,n0
subq n1,d,n1
.LLp6: cmpult n1,d,tmp
bne tmp,.LLp7
addq n0,1,n0
subq n1,d,n1
.LLp7:
stq n1,0(rem_ptr)
bis $31,n0,$0
ret $31,($26),1
.end __udiv_qrnnd

310
mpi/config.links Normal file
View File

@ -0,0 +1,310 @@
# config.links - helper for ../configure
# Copyright (C) 1998, 1999, 2000, 2001 Free Software Foundation, Inc.
#
# This file is part of GnuPG.
#
# GnuPG is free software; you can redistribute it and/or modify
# it under the terms of the GNU General Public License as published by
# the Free Software Foundation; either version 2 of the License, or
# (at your option) any later version.
#
# GnuPG is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with this program; if not, write to the Free Software
# Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA
# sourced by ../configure to get the list of files to link
# this should set $mpi_ln_list.
# Note: this is called from the above directory.
mpi_extra_modules=
mpi_sflags=
test -d ./mpi || mkdir ./mpi
echo '/* created by config.links - do not edit */' >./mpi/asm-syntax.h
if test "$try_asm_modules" = "yes" ; then
case "${target}" in
i[3467]86*-*-freebsd*-elf | \
i[3467]86*-*-freebsd[3-9]* | \
i[3467]86*-*-freebsdelf* | \
i[3467]86*-*-netbsd* )
echo '#define ELF_SYNTAX' >>./mpi/asm-syntax.h
cat $srcdir/mpi/i386/syntax.h >>./mpi/asm-syntax.h
path="i386"
;;
i586*-*-freebsd*-elf | \
i586*-*-freebsd[3-9]* | \
i586*-*-freebsdelf* | \
i586*-*-netbsd* | \
pentium-*-netbsd* | \
pentiumpro-*-netbsd*)
echo '#define ELF_SYNTAX' >>./mpi/asm-syntax.h
cat $srcdir/mpi/i386/syntax.h >>./mpi/asm-syntax.h
path="i586 i386"
;;
i[34]86*-*-bsdi4*)
echo '#define ELF_SYNTAX' >>./mpi/asm-syntax.h
cat $srcdir/mpi/i386/syntax.h >>./mpi/asm-syntax.h
path="i386"
;;
i[3467]86*-*-linuxaout* | \
i[3467]86*-*-linuxoldld* | \
i[3467]86*-*-*bsd*)
echo '#define BSD_SYNTAX' >>./mpi/asm-syntax.h
echo '#define X86_BROKEN_ALIGN' >>./mpi/asm-syntax.h
cat $srcdir/mpi/i386/syntax.h >>./mpi/asm-syntax.h
path="i386"
;;
i586*-*-linuxaout* | \
i586*-*-linuxoldld* | \
i586*-*-*bsd*)
echo '#define BSD_SYNTAX' >>./mpi/asm-syntax.h
echo '#define X86_BROKEN_ALIGN' >>./mpi/asm-syntax.h
cat $srcdir/mpi/i386/syntax.h >>./mpi/asm-syntax.h
path="i586 i386"
;;
i[3467]86*-msdosdjgpp*)
echo '#define BSD_SYNTAX' >>./mpi/asm-syntax.h
cat $srcdir/mpi/i386/syntax.h >>./mpi/asm-syntax.h
path="i386"
;;
i586*-msdosdjgpp*)
echo '#define BSD_SYNTAX' >>./mpi/asm-syntax.h
cat $srcdir/mpi/i386/syntax.h >>./mpi/asm-syntax.h
path="i586 i386"
;;
i[3467]86*-*-*)
echo '#define ELF_SYNTAX' >>./mpi/asm-syntax.h
cat $srcdir/mpi/i386/syntax.h >>./mpi/asm-syntax.h
path="i386"
;;
i586*-*-* | \
pentium-*-* | \
pentiumpro-*-*)
echo '#define ELF_SYNTAX' >>./mpi/asm-syntax.h
cat $srcdir/mpi/i386/syntax.h >>./mpi/asm-syntax.h
path="i586 i386"
;;
alpha*-*-*)
echo '/* configured for alpha */' >>./mpi/asm-syntax.h
path="alpha"
mpi_extra_modules="udiv-qrnnd"
;;
hppa7000*-*-*)
echo '/* configured for HPPA (pa7000) */' >>./mpi/asm-syntax.h
path="hppa1.1 hppa"
mpi_extra_modules="udiv-qrnnd"
;;
hppa1.0*-*-*)
echo '/* configured for HPPA 1.0 */' >>./mpi/asm-syntax.h
path="hppa"
mpi_extra_modules="udiv-qrnnd"
;;
hppa*-*-*) # assume pa7100
echo '/* configured for HPPA (pa7100) */' >>./mpi/asm-syntax.h
path="pa7100 hppa1.1 hppa"
mpi_extra_modules="udiv-qrnnd"
;;
sparc64-*-linux-gnu)
# An extra rule because we have an report for this one only.
# Should be compared against the next GMP version
echo '/* configured for sparc64-*-linux-gnu */' >>./mpi/asm-syntax.h
path="sparc32v8 sparc32"
mpi_extra_modules="udiv"
;;
sparc64-sun-solaris2*)
# Got a report that udiv is missing, so we try this one
echo '/* configured for sparc64-sun-solaris2 */' >>./mpi/asm-syntax.h
path="sparc32v8 sparc32"
mpi_extra_modules="udiv"
;;
sparc9*-*-* | \
sparc64*-*-* | \
ultrasparc*-*-* )
echo '/* configured for sparc9 or higher */' >>./mpi/asm-syntax.h
path="sparc32v8 sparc32"
;;
sparc8*-*-* | \
microsparc*-*-*)
echo '/* configured for sparc8 */' >>./mpi/asm-syntax.h
path="sparc32v8 sparc32"
;;
supersparc*-*-*)
echo '/* configured for supersparc */' >>./mpi/asm-syntax.h
path="supersparc sparc32v8 sparc32"
mpi_extra_modules="udiv"
;;
sparc*-*-*)
echo '/* configured for sparc */' >>./mpi/asm-syntax.h
path="sparc32"
mpi_extra_modules="udiv"
;;
mips[34]*-*-* | \
mips*-*-irix6*)
echo '/* configured for MIPS3 */' >>./mpi/asm-syntax.h
path="mips3"
;;
mips*-*-*)
echo '/* configured for MIPS2 */' >>./mpi/asm-syntax.h
path="mips2"
;;
# Motorola 68k configurations. Let m68k mean 68020-68040.
# mc68000 or mc68060 configurations need to be specified explicitly
m680[234]0*-*-linuxaout* | \
m68k*-*-linuxaout*)
echo '#define MIT_SYNTAX' >>./mpi/asm-syntax.h
cat $srcdir/mpi/m68k/syntax.h >>./mpi/asm-syntax.h
path="m68k/mc68020 m68k"
;;
m68060*-*-linuxaout*)
echo '#define MIT_SYNTAX' >>./mpi/asm-syntax.h
cat $srcdir/mpi/m68k/syntax.h >>./mpi/asm-syntax.h
path="m68k"
;;
m680[234]0*-*-linux* | \
m68k*-*-linux*)
echo '#define ELF_SYNTAX' >>./mpi/asm-syntax.h
cat $srcdir/mpi/m68k/syntax.h >>./mpi/asm-syntax.h
;;
m68060*-*-linux*)
echo '#define ELF_SYNTAX' >>./mpi/asm-syntax.h
cat $srcdir/mpi/m68k/syntax.h >>./mpi/asm-syntax.h
path="m68k"
;;
m68k-atari-mint)
echo '#define MIT_SYNTAX' >>./mpi/asm-syntax.h
cat $srcdir/mpi/m68k/syntax.h >>./mpi/asm-syntax.h
path="m68k/mc68020 m68k"
;;
m68000*-*-* | \
m68060*-*-*)
echo '#define MIT_SYNTAX' >>./mpi/asm-syntax.h
cat $srcdir/mpi/m68k/syntax.h >>./mpi/asm-syntax.h
path="m68k/mc68000"
;;
m680[234]0*-*-* | \
m68k*-*-*)
echo '#define MIT_SYNTAX' >>./mpi/asm-syntax.h
cat $srcdir/mpi/m68k/syntax.h >>./mpi/asm-syntax.h
path="m68k/mc68020 m68k"
;;
powerpc*-*-linux*)
echo '/* configured for powerpc/ELF */' >>./mpi/asm-syntax.h
echo '#define ELF_SYNTAX' >>./mpi/asm-syntax.h
cat $srcdir/mpi/powerpc32/syntax.h >>./mpi/asm-syntax.h
path="powerpc32"
;;
powerpc*-*-netbsd* | powerpc*-*-openbsd*)
echo '/* configured {Open,Net}BSD on powerpc */' >>./mpi/asm-syntax.h
echo '#define ELF_SYNTAX' >>./mpi/asm-syntax.h
cat $srcdir/mpi/powerpc32/syntax.h >>./mpi/asm-syntax.h
mpi_sflags="-Wa,-mppc"
path="powerpc32"
;;
rs6000-*-aix[456789]* | \
rs6000-*-aix3.2.[456789])
mpi_sflags="-Wa,-mpwr"
path="power"
mpi_extra_modules="udiv-w-sdiv"
;;
rs6000-*-* | \
power-*-* | \
power2-*-*)
mpi_sflags="-Wa,-mppc"
path="power"
mpi_extra_modules="udiv-w-sdiv"
;;
powerpc-ibm-aix4.2.* )
# I am not sure about this one but a machine identified by
# powerpc-ibm-aix4.2.1.0 cannot use the powerpc32 code.
mpi_sflags="-Wa,-mpwr"
path="power"
mpi_extra_modules="udiv-w-sdiv"
;;
ppc601-*-*)
mpi_sflags="-Wa,-mppc"
path="power powerpc32"
;;
ppc60[234]*-*-* | \
powerpc*-*-*)
mpi_sflags="-Wa,-mppc"
path="powerpc32"
;;
ppc620-*-* | \
powerpc64*-*-*)
mpi_sflags="-Wa,-mppc"
path="powerpc64"
;;
*)
echo '/* No assembler modules configured */' >>./mpi/asm-syntax.h
path=""
;;
esac
else
echo '/* Assembler modules disabled on request */' >>./mpi/asm-syntax.h
path=""
fi
# Make sysdep.h
echo '/* created by config.links - do not edit */' >./mpi/sysdep.h
if test x$ac_cv_sys_symbol_underscore = xyes; then
cat <<EOF >>./mpi/sysdep.h
#if __STDC__
#define C_SYMBOL_NAME(name) _##name
#else
#define C_SYMBOL_NAME(name) _/**/name
#endif
EOF
else
cat <<EOF >>./mpi/sysdep.h
#define C_SYMBOL_NAME(name) name
EOF
fi
# fixme: grep these modules from Makefile.in
mpi_ln_modules="${mpi_extra_modules} mpih-add1 mpih-mul1 mpih-mul2 mpih-mul3 \
mpih-lshift mpih-rshift mpih-sub1"
mpi_ln_objects=
mpi_ln_list=
# try to get file to link from the assembler subdirectory and
# if this fails get it from the generic subdirectory.
path="$path generic"
for fn in $mpi_ln_modules ; do
mpi_ln_objects="$mpi_ln_objects $fn.o"
for dir in $path ; do
rm -f $srcdir/mpi/$fn.[Sc]
if test -f $srcdir/mpi/$dir/$fn.S ; then
mpi_ln_list="$mpi_ln_list mpi/$fn.S:mpi/$dir/$fn.S"
break;
elif test -f $srcdir/mpi/$dir/$fn.c ; then
mpi_ln_list="$mpi_ln_list mpi/$fn.c:mpi/$dir/$fn.c"
break;
fi
done
done
# Same thing for the file which defines the limb size
path="$path generic"
for dir in $path ; do
rm -f $srcdir/mpi/mpi-asm-defs.h
if test -f $srcdir/mpi/$dir/mpi-asm-defs.h ; then
mpi_ln_list="$mpi_ln_list mpi/mpi-asm-defs.h:mpi/$dir/mpi-asm-defs.h"
break;
fi
done

95
mpi/g10m.c Normal file
View File

@ -0,0 +1,95 @@
/* g10m.c - Wrapper for MPI
* Copyright (C) 1998 Free Software Foundation, Inc.
*
* This file is part of GnuPG.
*
* GnuPG is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* GnuPG is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA
*/
#include <config.h>
#include <stdio.h>
#include <stdlib.h>
#include "mpi.h"
#include "util.h"
/* FIXME: The modules should use functions from libgcrypt */
const char *g10m_revision_string(int dummy) { return "$Revision$"; }
MPI
g10m_new( unsigned nbits )
{
return mpi_alloc( (nbits+BITS_PER_MPI_LIMB-1) / BITS_PER_MPI_LIMB );
}
MPI
g10m_new_secure( unsigned nbits )
{
return mpi_alloc_secure( (nbits+BITS_PER_MPI_LIMB-1) / BITS_PER_MPI_LIMB );
}
void
g10m_release( MPI a )
{
mpi_free(a);
}
void
g10m_resize( MPI a, unsigned nbits )
{
mpi_resize( a, (nbits+BITS_PER_MPI_LIMB-1) / BITS_PER_MPI_LIMB );
}
MPI g10m_copy( MPI a ) { return mpi_copy( a ); }
void g10m_swap( MPI a, MPI b) { mpi_swap( a, b ); }
void g10m_set( MPI w, MPI u) { mpi_set( w, u ); }
void g10m_set_ui( MPI w, ulong u ) { mpi_set_ui( w, u ); }
int g10m_cmp( MPI u, MPI v ) { return mpi_cmp( u, v ); }
int g10m_cmp_ui( MPI u, ulong v ) { return mpi_cmp_ui( u, v ); }
void g10m_add(MPI w, MPI u, MPI v) { mpi_add( w, u, v ); }
void g10m_add_ui(MPI w, MPI u, ulong v ) { mpi_add_ui( w, u, v ); }
void g10m_sub( MPI w, MPI u, MPI v) { mpi_sub( w, u, v ); }
void g10m_sub_ui(MPI w, MPI u, ulong v ) { mpi_sub_ui( w, u, v ); }
void g10m_mul( MPI w, MPI u, MPI v) { mpi_mul( w, u, v ); }
void g10m_mulm( MPI w, MPI u, MPI v, MPI m) { mpi_mulm( w, u, v, m ); }
void g10m_mul_2exp( MPI w, MPI u, ulong cnt) { mpi_mul_2exp( w, u, cnt ); }
void g10m_mul_ui(MPI w, MPI u, ulong v ) { mpi_mul_ui( w, u, v ); }
void g10m_fdiv_q( MPI q, MPI d, MPI r ) { mpi_fdiv_q( q, d, r ); }
void g10m_powm( MPI r, MPI b, MPI e, MPI m) { mpi_powm( r, b, e, m ); }
int g10m_gcd( MPI g, MPI a, MPI b ) { return mpi_gcd( g, a, b ); }
int g10m_invm( MPI x, MPI u, MPI v ) { mpi_invm( x, u, v ); return 0; }
unsigned g10m_get_nbits( MPI a ) { return mpi_get_nbits( a ); }
unsigned
g10m_get_size( MPI a )
{
return mpi_get_nlimbs( a ) * BITS_PER_MPI_LIMB;
}
void
g10m_set_buffer( MPI a, const char *buffer, unsigned nbytes, int sign )
{
mpi_set_buffer( a, buffer, nbytes, sign );
}

10
mpi/generic/distfiles Normal file
View File

@ -0,0 +1,10 @@
mpih-add1.c
mpih-mul1.c
mpih-mul2.c
mpih-mul3.c
mpih-lshift.c
mpih-rshift.c
mpih-sub1.c
udiv-w-sdiv.c
mpi-asm-defs.h

10
mpi/generic/mpi-asm-defs.h Normal file
View File

@ -0,0 +1,10 @@
/* This file defines some basic constants for the MPI machinery. We
* need to define the types on a per-CPU basis, so it is done with
* this file here. */
#define BYTES_PER_MPI_LIMB (SIZEOF_UNSIGNED_LONG)

65
mpi/generic/mpih-add1.c Normal file
View File

@ -0,0 +1,65 @@
/* mpihelp-add_1.c - MPI helper functions
* Copyright (C) 1994, 1996, 1997, 1998,
* 2000 Free Software Foundation, Inc.
*
* This file is part of GnuPG.
*
* GnuPG is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* GnuPG is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA
*
* Note: This code is heavily based on the GNU MP Library.
* Actually it's the same code with only minor changes in the
* way the data is stored; this is to support the abstraction
* of an optional secure memory allocation which may be used
* to avoid revealing of sensitive data due to paging etc.
* The GNU MP Library itself is published under the LGPL;
* however I decided to publish this code under the plain GPL.
*/
#include <config.h>
#include <stdio.h>
#include <stdlib.h>
#include "mpi-internal.h"
#include "longlong.h"
mpi_limb_t
mpihelp_add_n( mpi_ptr_t res_ptr, mpi_ptr_t s1_ptr,
mpi_ptr_t s2_ptr, mpi_size_t size)
{
mpi_limb_t x, y, cy;
mpi_size_t j;
/* The loop counter and index J goes from -SIZE to -1. This way
the loop becomes faster. */
j = -size;
/* Offset the base pointers to compensate for the negative indices. */
s1_ptr -= j;
s2_ptr -= j;
res_ptr -= j;
cy = 0;
do {
y = s2_ptr[j];
x = s1_ptr[j];
y += cy; /* add previous carry to one addend */
cy = y < cy; /* get out carry from that addition */
y += x; /* add other addend */
cy += y < x; /* get out carry from that add, combine */
res_ptr[j] = y;
} while( ++j );
return cy;
}

69
mpi/generic/mpih-lshift.c Normal file
View File

@ -0,0 +1,69 @@
/* mpihelp-lshift.c - MPI helper functions
* Copyright (C) 1994, 1996, 1998, 2001 Free Software Foundation, Inc.
*
* This file is part of GnuPG.
*
* GnuPG is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* GnuPG is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA
*
* Note: This code is heavily based on the GNU MP Library.
* Actually it's the same code with only minor changes in the
* way the data is stored; this is to support the abstraction
* of an optional secure memory allocation which may be used
* to avoid revealing of sensitive data due to paging etc.
* The GNU MP Library itself is published under the LGPL;
* however I decided to publish this code under the plain GPL.
*/
#include <config.h>
#include <stdio.h>
#include <stdlib.h>
#include "mpi-internal.h"
/* Shift U (pointed to by UP and USIZE digits long) CNT bits to the left
* and store the USIZE least significant digits of the result at WP.
* Return the bits shifted out from the most significant digit.
*
* Argument constraints:
* 1. 0 < CNT < BITS_PER_MP_LIMB
* 2. If the result is to be written over the input, WP must be >= UP.
*/
mpi_limb_t
mpihelp_lshift( mpi_ptr_t wp, mpi_ptr_t up, mpi_size_t usize,
unsigned int cnt)
{
mpi_limb_t high_limb, low_limb;
unsigned sh_1, sh_2;
mpi_size_t i;
mpi_limb_t retval;
sh_1 = cnt;
wp += 1;
sh_2 = BITS_PER_MPI_LIMB - sh_1;
i = usize - 1;
low_limb = up[i];
retval = low_limb >> sh_2;
high_limb = low_limb;
while( --i >= 0 ) {
low_limb = up[i];
wp[i] = (high_limb << sh_1) | (low_limb >> sh_2);
high_limb = low_limb;
}
wp[i] = high_limb << sh_1;
return retval;
}

61
mpi/generic/mpih-mul1.c Normal file
View File

@ -0,0 +1,61 @@
/* mpihelp-mul_1.c - MPI helper functions
* Copyright (C) 1994, 1996, 1997, 1998, 2001 Free Software Foundation, Inc.
*
* This file is part of GnuPG.
*
* GnuPG is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* GnuPG is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA
*
* Note: This code is heavily based on the GNU MP Library.
* Actually it's the same code with only minor changes in the
* way the data is stored; this is to support the abstraction
* of an optional secure memory allocation which may be used
* to avoid revealing of sensitive data due to paging etc.
* The GNU MP Library itself is published under the LGPL;
* however I decided to publish this code under the plain GPL.
*/
#include <config.h>
#include <stdio.h>
#include <stdlib.h>
#include "mpi-internal.h"
#include "longlong.h"
mpi_limb_t
mpihelp_mul_1( mpi_ptr_t res_ptr, mpi_ptr_t s1_ptr, mpi_size_t s1_size,
mpi_limb_t s2_limb)
{
mpi_limb_t cy_limb;
mpi_size_t j;
mpi_limb_t prod_high, prod_low;
/* The loop counter and index J goes from -S1_SIZE to -1. This way
* the loop becomes faster. */
j = -s1_size;
/* Offset the base pointers to compensate for the negative indices. */
s1_ptr -= j;
res_ptr -= j;
cy_limb = 0;
do {
umul_ppmm( prod_high, prod_low, s1_ptr[j], s2_limb );
prod_low += cy_limb;
cy_limb = (prod_low < cy_limb?1:0) + prod_high;
res_ptr[j] = prod_low;
} while( ++j );
return cy_limb;
}

66
mpi/generic/mpih-mul2.c Normal file
View File

@ -0,0 +1,66 @@
/* mpihelp-mul_2.c - MPI helper functions
* Copyright (C) 1994, 1996, 1997, 1998, 2001 Free Software Foundation, Inc.
*
* This file is part of GnuPG.
*
* GnuPG is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* GnuPG is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA
*
* Note: This code is heavily based on the GNU MP Library.
* Actually it's the same code with only minor changes in the
* way the data is stored; this is to support the abstraction
* of an optional secure memory allocation which may be used
* to avoid revealing of sensitive data due to paging etc.
* The GNU MP Library itself is published under the LGPL;
* however I decided to publish this code under the plain GPL.
*/
#include <config.h>
#include <stdio.h>
#include <stdlib.h>
#include "mpi-internal.h"
#include "longlong.h"
mpi_limb_t
mpihelp_addmul_1( mpi_ptr_t res_ptr, mpi_ptr_t s1_ptr,
mpi_size_t s1_size, mpi_limb_t s2_limb)
{
mpi_limb_t cy_limb;
mpi_size_t j;
mpi_limb_t prod_high, prod_low;
mpi_limb_t x;
/* The loop counter and index J goes from -SIZE to -1. This way
* the loop becomes faster. */
j = -s1_size;
res_ptr -= j;
s1_ptr -= j;
cy_limb = 0;
do {
umul_ppmm( prod_high, prod_low, s1_ptr[j], s2_limb );
prod_low += cy_limb;
cy_limb = (prod_low < cy_limb?1:0) + prod_high;
x = res_ptr[j];
prod_low = x + prod_low;
cy_limb += prod_low < x?1:0;
res_ptr[j] = prod_low;
} while ( ++j );
return cy_limb;
}

67
mpi/generic/mpih-mul3.c Normal file
View File

@ -0,0 +1,67 @@
/* mpihelp-mul_3.c - MPI helper functions
* Copyright (C) 1994, 1996, 1997, 1998, 2001 Free Software Foundation, Inc.
*
* This file is part of GnuPG.
*
* GnuPG is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* GnuPG is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA
*
* Note: This code is heavily based on the GNU MP Library.
* Actually it's the same code with only minor changes in the
* way the data is stored; this is to support the abstraction
* of an optional secure memory allocation which may be used
* to avoid revealing of sensitive data due to paging etc.
* The GNU MP Library itself is published under the LGPL;
* however I decided to publish this code under the plain GPL.
*/
#include <config.h>
#include <stdio.h>
#include <stdlib.h>
#include "mpi-internal.h"
#include "longlong.h"
mpi_limb_t
mpihelp_submul_1( mpi_ptr_t res_ptr, mpi_ptr_t s1_ptr,
mpi_size_t s1_size, mpi_limb_t s2_limb)
{
mpi_limb_t cy_limb;
mpi_size_t j;
mpi_limb_t prod_high, prod_low;
mpi_limb_t x;
/* The loop counter and index J goes from -SIZE to -1. This way
* the loop becomes faster. */
j = -s1_size;
res_ptr -= j;
s1_ptr -= j;
cy_limb = 0;
do {
umul_ppmm( prod_high, prod_low, s1_ptr[j], s2_limb);
prod_low += cy_limb;
cy_limb = (prod_low < cy_limb?1:0) + prod_high;
x = res_ptr[j];
prod_low = x - prod_low;
cy_limb += prod_low > x?1:0;
res_ptr[j] = prod_low;
} while( ++j );
return cy_limb;
}

68
mpi/generic/mpih-rshift.c Normal file
View File

@ -0,0 +1,68 @@
/* mpih-rshift.c - MPI helper functions
* Copyright (C) 1994, 1996, 1998, 1999,
* 2000, 2001 Free Software Foundation, Inc.
*
* This file is part of GNUPG
*
* GNUPG is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* GNUPG is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA
*
* Note: This code is heavily based on the GNU MP Library.
* Actually it's the same code with only minor changes in the
* way the data is stored; this is to support the abstraction
* of an optional secure memory allocation which may be used
* to avoid revealing of sensitive data due to paging etc.
* The GNU MP Library itself is published under the LGPL;
* however I decided to publish this code under the plain GPL.
*/
#include <config.h>
#include <stdio.h>
#include <stdlib.h>
#include "mpi-internal.h"
/* Shift U (pointed to by UP and USIZE limbs long) CNT bits to the right
* and store the USIZE least significant limbs of the result at WP.
* The bits shifted out to the right are returned.
*
* Argument constraints:
* 1. 0 < CNT < BITS_PER_MP_LIMB
* 2. If the result is to be written over the input, WP must be <= UP.
*/
mpi_limb_t
mpihelp_rshift( mpi_ptr_t wp, mpi_ptr_t up, mpi_size_t usize, unsigned cnt)
{
mpi_limb_t high_limb, low_limb;
unsigned sh_1, sh_2;
mpi_size_t i;
mpi_limb_t retval;
sh_1 = cnt;
wp -= 1;
sh_2 = BITS_PER_MPI_LIMB - sh_1;
high_limb = up[0];
retval = high_limb << sh_2;
low_limb = high_limb;
for( i=1; i < usize; i++) {
high_limb = up[i];
wp[i] = (low_limb >> sh_1) | (high_limb << sh_2);
low_limb = high_limb;
}
wp[i] = low_limb >> sh_1;
return retval;
}

65
mpi/generic/mpih-sub1.c Normal file
View File

@ -0,0 +1,65 @@
/* mpihelp-add_2.c - MPI helper functions
* Copyright (C) 1994, 1996, 1997, 1998, 2001 Free Software Foundation, Inc.
*
* This file is part of GnuPG.
*
* GnuPG is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* GnuPG is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA
*
* Note: This code is heavily based on the GNU MP Library.
* Actually it's the same code with only minor changes in the
* way the data is stored; this is to support the abstraction
* of an optional secure memory allocation which may be used
* to avoid revealing of sensitive data due to paging etc.
* The GNU MP Library itself is published under the LGPL;
* however I decided to publish this code under the plain GPL.
*/
#include <config.h>
#include <stdio.h>
#include <stdlib.h>
#include "mpi-internal.h"
#include "longlong.h"
mpi_limb_t
mpihelp_sub_n( mpi_ptr_t res_ptr, mpi_ptr_t s1_ptr,
mpi_ptr_t s2_ptr, mpi_size_t size)
{
mpi_limb_t x, y, cy;
mpi_size_t j;
/* The loop counter and index J goes from -SIZE to -1. This way
the loop becomes faster. */
j = -size;
/* Offset the base pointers to compensate for the negative indices. */
s1_ptr -= j;
s2_ptr -= j;
res_ptr -= j;
cy = 0;
do {
y = s2_ptr[j];
x = s1_ptr[j];
y += cy; /* add previous carry to subtrahend */
cy = y < cy; /* get out carry from that addition */
y = x - y; /* main subtract */
cy += y > x; /* get out carry from the subtract, combine */
res_ptr[j] = y;
} while( ++j );
return cy;
}

133
mpi/generic/udiv-w-sdiv.c Normal file
View File

@ -0,0 +1,133 @@
/* mpihelp_udiv_w_sdiv -- implement udiv_qrnnd on machines with only signed
* division.
* Copyright (C) 1992, 1994, 1996, 1998 Free Software Foundation, Inc.
* Contributed by Peter L. Montgomery.
*
* This file is part of GnuPG.
*
* GnuPG is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* GnuPG is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA
*/
#include <config.h>
#include <stdio.h>
#include <stdlib.h>
#include "mpi-internal.h"
#include "longlong.h"
#if 0 /* not yet ported to MPI */
mpi_limb_t
mpihelp_udiv_w_sdiv( mpi_limp_t *rp,
mpi_limp_t *a1,
mpi_limp_t *a0,
mpi_limp_t *d )
{
mp_limb_t q, r;
mp_limb_t c0, c1, b1;
if ((mpi_limb_signed_t) d >= 0)
{
if (a1 < d - a1 - (a0 >> (BITS_PER_MP_LIMB - 1)))
{
/* dividend, divisor, and quotient are nonnegative */
sdiv_qrnnd (q, r, a1, a0, d);
}
else
{
/* Compute c1*2^32 + c0 = a1*2^32 + a0 - 2^31*d */
sub_ddmmss (c1, c0, a1, a0, d >> 1, d << (BITS_PER_MP_LIMB - 1));
/* Divide (c1*2^32 + c0) by d */
sdiv_qrnnd (q, r, c1, c0, d);
/* Add 2^31 to quotient */
q += (mp_limb_t) 1 << (BITS_PER_MP_LIMB - 1);
}
}
else
{
b1 = d >> 1; /* d/2, between 2^30 and 2^31 - 1 */
c1 = a1 >> 1; /* A/2 */
c0 = (a1 << (BITS_PER_MP_LIMB - 1)) + (a0 >> 1);
if (a1 < b1) /* A < 2^32*b1, so A/2 < 2^31*b1 */
{
sdiv_qrnnd (q, r, c1, c0, b1); /* (A/2) / (d/2) */
r = 2*r + (a0 & 1); /* Remainder from A/(2*b1) */
if ((d & 1) != 0)
{
if (r >= q)
r = r - q;
else if (q - r <= d)
{
r = r - q + d;
q--;
}
else
{
r = r - q + 2*d;
q -= 2;
}
}
}
else if (c1 < b1) /* So 2^31 <= (A/2)/b1 < 2^32 */
{
c1 = (b1 - 1) - c1;
c0 = ~c0; /* logical NOT */
sdiv_qrnnd (q, r, c1, c0, b1); /* (A/2) / (d/2) */
q = ~q; /* (A/2)/b1 */
r = (b1 - 1) - r;
r = 2*r + (a0 & 1); /* A/(2*b1) */
if ((d & 1) != 0)
{
if (r >= q)
r = r - q;
else if (q - r <= d)
{
r = r - q + d;
q--;
}
else
{
r = r - q + 2*d;
q -= 2;
}
}
}
else /* Implies c1 = b1 */
{ /* Hence a1 = d - 1 = 2*b1 - 1 */
if (a0 >= -d)
{
q = -1;
r = a0 + d;
}
else
{
q = -2;
r = a0 + 2*d;
}
}
}
*rp = r;
return q;
}
#endif

84
mpi/hppa/README Normal file
View File

@ -0,0 +1,84 @@
This directory contains mpn functions for various HP PA-RISC chips. Code
that runs faster on the PA7100 and later implementations, is in the pa7100
directory.
RELEVANT OPTIMIZATION ISSUES
Load and Store timing
On the PA7000 no memory instructions can issue the two cycles after a store.
For the PA7100, this is reduced to one cycle.
The PA7100 has a lookup-free cache, so it helps to schedule loads and the
dependent instruction really far from each other.
STATUS
1. mpn_mul_1 could be improved to 6.5 cycles/limb on the PA7100, using the
instructions bwlow (but some sw pipelining is needed to avoid the
xmpyu-fstds delay):
fldds s1_ptr
xmpyu
fstds N(%r30)
xmpyu
fstds N(%r30)
ldws N(%r30)
ldws N(%r30)
ldws N(%r30)
ldws N(%r30)
addc
stws res_ptr
addc
stws res_ptr
addib Loop
2. mpn_addmul_1 could be improved from the current 10 to 7.5 cycles/limb
(asymptotically) on the PA7100, using the instructions below. With proper
sw pipelining and the unrolling level below, the speed becomes 8
cycles/limb.
fldds s1_ptr
fldds s1_ptr
xmpyu
fstds N(%r30)
xmpyu
fstds N(%r30)
xmpyu
fstds N(%r30)
xmpyu
fstds N(%r30)
ldws N(%r30)
ldws N(%r30)
ldws N(%r30)
ldws N(%r30)
ldws N(%r30)
ldws N(%r30)
ldws N(%r30)
ldws N(%r30)
addc
addc
addc
addc
addc %r0,%r0,cy-limb
ldws res_ptr
ldws res_ptr
ldws res_ptr
ldws res_ptr
add
stws res_ptr
addc
stws res_ptr
addc
stws res_ptr
addc
stws res_ptr
addib

7
mpi/hppa/distfiles Normal file
View File

@ -0,0 +1,7 @@
README
udiv-qrnnd.S
mpih-add1.S
mpih-sub1.S
mpih-lshift.S
mpih-rshift.S

72
mpi/hppa/mpih-add1.S Normal file
View File

@ -0,0 +1,72 @@
/* hppa add_n -- Add two limb vectors of the same length > 0 and store
* sum in a third limb vector.
*
* Copyright (C) 1992, 1994, 1998,
* 2001 Fee Software Foundation, Inc.
*
* This file is part of GnuPG.
*
* GnuPG is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* GnuPG is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA
*
* Note: This code is heavily based on the GNU MP Library.
* Actually it's the same code with only minor changes in the
* way the data is stored; this is to support the abstraction
* of an optional secure memory allocation which may be used
* to avoid revealing of sensitive data due to paging etc.
* The GNU MP Library itself is published under the LGPL;
* however I decided to publish this code under the plain GPL.
*/
/*******************
* mpi_limb_t
* mpihelp_add_n( mpi_ptr_t res_ptr, (gr26)
* mpi_ptr_t s1_ptr, (gr25)
* mpi_ptr_t s2_ptr, (gr24)
* mpi_size_t size) (gr23)
*
* One might want to unroll this as for other processors, but it turns
* out that the data cache contention after a store makes such
* unrolling useless. We can't come under 5 cycles/limb anyway.
*/
.code
.export mpihelp_add_n
.label mpihelp_add_n
.proc
.callinfo frame=0,no_calls
.entry
ldws,ma 4(0,%r25),%r20
ldws,ma 4(0,%r24),%r19
addib,= -1,%r23,L$end ; check for (SIZE == 1)
add %r20,%r19,%r28 ; add first limbs ignoring cy
.label L$loop
ldws,ma 4(0,%r25),%r20
ldws,ma 4(0,%r24),%r19
stws,ma %r28,4(0,%r26)
addib,<> -1,%r23,L$loop
addc %r20,%r19,%r28
.label L$end
stws %r28,0(0,%r26)
bv 0(%r2)
addc %r0,%r0,%r28
.exit
.procend

77
mpi/hppa/mpih-lshift.S Normal file
View File

@ -0,0 +1,77 @@
/* hppa lshift
*
* Copyright (C) 1992, 1994, 1998
* 2001 Free Software Foundation, Inc.
*
* This file is part of GnuPG.
*
* GnuPG is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* GnuPG is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA
*/
/*******************
* mpi_limb_t
* mpihelp_lshift( mpi_ptr_t wp, (gr26)
* mpi_ptr_t up, (gr25)
* mpi_size_t usize, (gr24)
* unsigned cnt) (gr23)
*/
.code
.export mpihelp_lshift
.label mpihelp_lshift
.proc
.callinfo frame=64,no_calls
.entry
sh2add %r24,%r25,%r25
sh2add %r24,%r26,%r26
ldws,mb -4(0,%r25),%r22
subi 32,%r23,%r1
mtsar %r1
addib,= -1,%r24,L$0004
vshd %r0,%r22,%r28 ; compute carry out limb
ldws,mb -4(0,%r25),%r29
addib,= -1,%r24,L$0002
vshd %r22,%r29,%r20
.label L$loop
ldws,mb -4(0,%r25),%r22
stws,mb %r20,-4(0,%r26)
addib,= -1,%r24,L$0003
vshd %r29,%r22,%r20
ldws,mb -4(0,%r25),%r29
stws,mb %r20,-4(0,%r26)
addib,<> -1,%r24,L$loop
vshd %r22,%r29,%r20
.label L$0002
stws,mb %r20,-4(0,%r26)
vshd %r29,%r0,%r20
bv 0(%r2)
stw %r20,-4(0,%r26)
.label L$0003
stws,mb %r20,-4(0,%r26)
.label L$0004
vshd %r22,%r0,%r20
bv 0(%r2)
stw %r20,-4(0,%r26)
.exit
.procend

73
mpi/hppa/mpih-rshift.S Normal file
View File

@ -0,0 +1,73 @@
/* hppa rshift
*
* Copyright (C) 1992, 1994, 1998,
* 2001 Free Software Foundation, Inc.
*
* This file is part of GnuPG.
*
* GnuPG is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* GnuPG is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA
*/
/*******************
* mpi_limb_t
* mpihelp_rshift( mpi_ptr_t wp, (gr26)
* mpi_ptr_t up, (gr25)
* mpi_size_t usize, (gr24)
* unsigned cnt) (gr23)
*/
.code
.export mpihelp_rshift
.label mpihelp_rshift
.proc
.callinfo frame=64,no_calls
.entry
ldws,ma 4(0,%r25),%r22
mtsar %r23
addib,= -1,%r24,L$r004
vshd %r22,%r0,%r28 ; compute carry out limb
ldws,ma 4(0,%r25),%r29
addib,= -1,%r24,L$r002
vshd %r29,%r22,%r20
.label L$roop
ldws,ma 4(0,%r25),%r22
stws,ma %r20,4(0,%r26)
addib,= -1,%r24,L$r003
vshd %r22,%r29,%r20
ldws,ma 4(0,%r25),%r29
stws,ma %r20,4(0,%r26)
addib,<> -1,%r24,L$roop
vshd %r29,%r22,%r20
.label L$r002
stws,ma %r20,4(0,%r26)
vshd %r0,%r29,%r20
bv 0(%r2)
stw %r20,0(0,%r26)
.label L$r003
stws,ma %r20,4(0,%r26)
.label L$r004
vshd %r0,%r22,%r20
bv 0(%r2)
stw %r20,0(0,%r26)
.exit
.procend

80
mpi/hppa/mpih-sub1.S Normal file
View File

@ -0,0 +1,80 @@
/* hppa sub_n -- Sub two limb vectors of the same length > 0 and store
* sum in a third limb vector.
*
* Copyright (C) 1992, 1994, 1998,
* 2001 Free Software Foundation, Inc.
*
* This file is part of GnuPG.
*
* GnuPG is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* GnuPG is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA
*
* Note: This code is heavily based on the GNU MP Library.
* Actually it's the same code with only minor changes in the
* way the data is stored; this is to support the abstraction
* of an optional secure memory allocation which may be used
* to avoid revealing of sensitive data due to paging etc.
* The GNU MP Library itself is published under the LGPL;
* however I decided to publish this code under the plain GPL.
*/
#include "sysdep.h"
#include "asm-syntax.h"
/*******************
* mpi_limb_t
* mpihelp_sub_n( mpi_ptr_t res_ptr, (gr26)
* mpi_ptr_t s1_ptr, (gr25)
* mpi_ptr_t s2_ptr, (gr24)
* mpi_size_t size) (gr23)
*
* One might want to unroll this as for other processors, but it turns
* out that the data cache contention after a store makes such
* unrolling useless. We can't come under 5 cycles/limb anyway.
*/
.code
.export mpihelp_sub_n
.label mpihelp_sub_n
.proc
.callinfo frame=0,no_calls
.entry
ldws,ma 4(0,%r25),%r20
ldws,ma 4(0,%r24),%r19
addib,= -1,%r23,L$end ; check for (SIZE == 1)
sub %r20,%r19,%r28 ; subtract first limbs ignoring cy
.label L$loop
ldws,ma 4(0,%r25),%r20
ldws,ma 4(0,%r24),%r19
stws,ma %r28,4(0,%r26)
addib,<> -1,%r23,L$loop
subb %r20,%r19,%r28
.label L$end
stws %r28,0(0,%r26)
addc %r0,%r0,%r28
bv 0(%r2)
subi 1,%r28,%r28
.exit
.procend

298
mpi/hppa/udiv-qrnnd.S Normal file
View File

@ -0,0 +1,298 @@
/* HP-PA __udiv_qrnnd division support, used from longlong.h.
* This version runs fast on pre-PA7000 CPUs.
*
* Copyright (C) 1993, 1994, 1998, 2001 Free Software Foundation, Inc.
*
* This file is part of GnuPG.
*
* GnuPG is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* GnuPG is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA
*
* Note: This code is heavily based on the GNU MP Library.
* Actually it's the same code with only minor changes in the
* way the data is stored; this is to support the abstraction
* of an optional secure memory allocation which may be used
* to avoid revealing of sensitive data due to paging etc.
* The GNU MP Library itself is published under the LGPL;
* however I decided to publish this code under the plain GPL.
*/
/* INPUT PARAMETERS
* rem_ptr gr26
* n1 gr25
* n0 gr24
* d gr23
*
* The code size is a bit excessive. We could merge the last two ds;addc
* sequences by simply moving the "bb,< Odd" instruction down. The only
* trouble is the FFFFFFFF code that would need some hacking.
*/
.code
.export __udiv_qrnnd
.label __udiv_qrnnd
.proc
.callinfo frame=0,no_calls
.entry
comb,< %r23,0,L$largedivisor
sub %r0,%r23,%r1 ; clear cy as side-effect
ds %r0,%r1,%r0
addc %r24,%r24,%r24
ds %r25,%r23,%r25
addc %r24,%r24,%r24
ds %r25,%r23,%r25
addc %r24,%r24,%r24
ds %r25,%r23,%r25
addc %r24,%r24,%r24
ds %r25,%r23,%r25
addc %r24,%r24,%r24
ds %r25,%r23,%r25
addc %r24,%r24,%r24
ds %r25,%r23,%r25
addc %r24,%r24,%r24
ds %r25,%r23,%r25
addc %r24,%r24,%r24
ds %r25,%r23,%r25
addc %r24,%r24,%r24
ds %r25,%r23,%r25
addc %r24,%r24,%r24
ds %r25,%r23,%r25
addc %r24,%r24,%r24
ds %r25,%r23,%r25
addc %r24,%r24,%r24
ds %r25,%r23,%r25
addc %r24,%r24,%r24
ds %r25,%r23,%r25
addc %r24,%r24,%r24
ds %r25,%r23,%r25
addc %r24,%r24,%r24
ds %r25,%r23,%r25
addc %r24,%r24,%r24
ds %r25,%r23,%r25
addc %r24,%r24,%r24
ds %r25,%r23,%r25
addc %r24,%r24,%r24
ds %r25,%r23,%r25
addc %r24,%r24,%r24
ds %r25,%r23,%r25
addc %r24,%r24,%r24
ds %r25,%r23,%r25
addc %r24,%r24,%r24
ds %r25,%r23,%r25
addc %r24,%r24,%r24
ds %r25,%r23,%r25
addc %r24,%r24,%r24
ds %r25,%r23,%r25
addc %r24,%r24,%r24
ds %r25,%r23,%r25
addc %r24,%r24,%r24
ds %r25,%r23,%r25
addc %r24,%r24,%r24
ds %r25,%r23,%r25
addc %r24,%r24,%r24
ds %r25,%r23,%r25
addc %r24,%r24,%r24
ds %r25,%r23,%r25
addc %r24,%r24,%r24
ds %r25,%r23,%r25
addc %r24,%r24,%r24
ds %r25,%r23,%r25
addc %r24,%r24,%r24
ds %r25,%r23,%r25
addc %r24,%r24,%r28
ds %r25,%r23,%r25
comclr,>= %r25,%r0,%r0
addl %r25,%r23,%r25
stws %r25,0(0,%r26)
bv 0(%r2)
addc %r28,%r28,%r28
.label L$largedivisor
extru %r24,31,1,%r19 ; r19 = n0 & 1
bb,< %r23,31,L$odd
extru %r23,30,31,%r22 ; r22 = d >> 1
shd %r25,%r24,1,%r24 ; r24 = new n0
extru %r25,30,31,%r25 ; r25 = new n1
sub %r0,%r22,%r21
ds %r0,%r21,%r0
addc %r24,%r24,%r24
ds %r25,%r22,%r25
addc %r24,%r24,%r24
ds %r25,%r22,%r25
addc %r24,%r24,%r24
ds %r25,%r22,%r25
addc %r24,%r24,%r24
ds %r25,%r22,%r25
addc %r24,%r24,%r24
ds %r25,%r22,%r25
addc %r24,%r24,%r24
ds %r25,%r22,%r25
addc %r24,%r24,%r24
ds %r25,%r22,%r25
addc %r24,%r24,%r24
ds %r25,%r22,%r25
addc %r24,%r24,%r24
ds %r25,%r22,%r25
addc %r24,%r24,%r24
ds %r25,%r22,%r25
addc %r24,%r24,%r24
ds %r25,%r22,%r25
addc %r24,%r24,%r24
ds %r25,%r22,%r25
addc %r24,%r24,%r24
ds %r25,%r22,%r25
addc %r24,%r24,%r24
ds %r25,%r22,%r25
addc %r24,%r24,%r24
ds %r25,%r22,%r25
addc %r24,%r24,%r24
ds %r25,%r22,%r25
addc %r24,%r24,%r24
ds %r25,%r22,%r25
addc %r24,%r24,%r24
ds %r25,%r22,%r25
addc %r24,%r24,%r24
ds %r25,%r22,%r25
addc %r24,%r24,%r24
ds %r25,%r22,%r25
addc %r24,%r24,%r24
ds %r25,%r22,%r25
addc %r24,%r24,%r24
ds %r25,%r22,%r25
addc %r24,%r24,%r24
ds %r25,%r22,%r25
addc %r24,%r24,%r24
ds %r25,%r22,%r25
addc %r24,%r24,%r24
ds %r25,%r22,%r25
addc %r24,%r24,%r24
ds %r25,%r22,%r25
addc %r24,%r24,%r24
ds %r25,%r22,%r25
addc %r24,%r24,%r24
ds %r25,%r22,%r25
addc %r24,%r24,%r24
ds %r25,%r22,%r25
addc %r24,%r24,%r24
ds %r25,%r22,%r25
addc %r24,%r24,%r24
ds %r25,%r22,%r25
addc %r24,%r24,%r24
ds %r25,%r22,%r25
comclr,>= %r25,%r0,%r0
addl %r25,%r22,%r25
sh1addl %r25,%r19,%r25
stws %r25,0(0,%r26)
bv 0(%r2)
addc %r24,%r24,%r28
.label L$odd
addib,sv,n 1,%r22,L$FF.. ; r22 = (d / 2 + 1)
shd %r25,%r24,1,%r24 ; r24 = new n0
extru %r25,30,31,%r25 ; r25 = new n1
sub %r0,%r22,%r21
ds %r0,%r21,%r0
addc %r24,%r24,%r24
ds %r25,%r22,%r25
addc %r24,%r24,%r24
ds %r25,%r22,%r25
addc %r24,%r24,%r24
ds %r25,%r22,%r25
addc %r24,%r24,%r24
ds %r25,%r22,%r25
addc %r24,%r24,%r24
ds %r25,%r22,%r25
addc %r24,%r24,%r24
ds %r25,%r22,%r25
addc %r24,%r24,%r24
ds %r25,%r22,%r25
addc %r24,%r24,%r24
ds %r25,%r22,%r25
addc %r24,%r24,%r24
ds %r25,%r22,%r25
addc %r24,%r24,%r24
ds %r25,%r22,%r25
addc %r24,%r24,%r24
ds %r25,%r22,%r25
addc %r24,%r24,%r24
ds %r25,%r22,%r25
addc %r24,%r24,%r24
ds %r25,%r22,%r25
addc %r24,%r24,%r24
ds %r25,%r22,%r25
addc %r24,%r24,%r24
ds %r25,%r22,%r25
addc %r24,%r24,%r24
ds %r25,%r22,%r25
addc %r24,%r24,%r24
ds %r25,%r22,%r25
addc %r24,%r24,%r24
ds %r25,%r22,%r25
addc %r24,%r24,%r24
ds %r25,%r22,%r25
addc %r24,%r24,%r24
ds %r25,%r22,%r25
addc %r24,%r24,%r24
ds %r25,%r22,%r25
addc %r24,%r24,%r24
ds %r25,%r22,%r25
addc %r24,%r24,%r24
ds %r25,%r22,%r25
addc %r24,%r24,%r24
ds %r25,%r22,%r25
addc %r24,%r24,%r24
ds %r25,%r22,%r25
addc %r24,%r24,%r24
ds %r25,%r22,%r25
addc %r24,%r24,%r24
ds %r25,%r22,%r25
addc %r24,%r24,%r24
ds %r25,%r22,%r25
addc %r24,%r24,%r24
ds %r25,%r22,%r25
addc %r24,%r24,%r24
ds %r25,%r22,%r25
addc %r24,%r24,%r24
ds %r25,%r22,%r25
addc %r24,%r24,%r24
ds %r25,%r22,%r25
addc %r24,%r24,%r28
comclr,>= %r25,%r0,%r0
addl %r25,%r22,%r25
sh1addl %r25,%r19,%r25
; We have computed (n1,,n0) / (d + 1), q' = r28, r' = r25
add,nuv %r28,%r25,%r25
addl %r25,%r1,%r25
addc %r0,%r28,%r28
sub,<< %r25,%r23,%r0
addl %r25,%r1,%r25
stws %r25,0(0,%r26)
bv 0(%r2)
addc %r0,%r28,%r28
; This is just a special case of the code above.
; We come here when d == 0xFFFFFFFF
.label L$FF..
add,uv %r25,%r24,%r24
sub,<< %r24,%r23,%r0
ldo 1(%r24),%r24
stws %r24,0(0,%r26)
bv 0(%r2)
addc %r0,%r25,%r28
.exit
.procend

5
mpi/hppa1.1/distfiles Normal file
View File

@ -0,0 +1,5 @@
udiv-qrnnd.S
mpih-mul1.S
mpih-mul2.S
mpih-mul3.S

115
mpi/hppa1.1/mpih-mul1.S Normal file
View File

@ -0,0 +1,115 @@
/* hppa1.1 mul_1 -- Multiply a limb vector with a limb and store
* the result in a second limb vector.
*
* Copyright (C) 1992, 1993, 1994, 1998,
* 2001 Free Software Foundation, Inc.
*
* This file is part of GnuPG.
*
* GnuPG is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* GnuPG is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA
*
* Note: This code is heavily based on the GNU MP Library.
* Actually it's the same code with only minor changes in the
* way the data is stored; this is to support the abstraction
* of an optional secure memory allocation which may be used
* to avoid revealing of sensitive data due to paging etc.
* The GNU MP Library itself is published under the LGPL;
* however I decided to publish this code under the plain GPL.
*/
/*******************
* mpi_limb_t
* mpihelp_mul_1( mpi_ptr_t res_ptr, (r26)
* mpi_ptr_t s1_ptr, (r25)
* mpi_size_t s1_size, (r24)
* mpi_limb_t s2_limb) (r23)
*
*
*
* This runs at 9 cycles/limb on a PA7000. With the used instructions, it can
* not become faster due to data cache contention after a store. On the
* PA7100 it runs at 7 cycles/limb, and that can not be improved either, since
* only the xmpyu does not need the integer pipeline, so the only dual-issue
* we will get are addc+xmpyu. Unrolling would not help either CPU.
*
* We could use fldds to read two limbs at a time from the S1 array, and that
* could bring down the times to 8.5 and 6.5 cycles/limb for the PA7000 and
* PA7100, respectively. We don't do that since it does not seem worth the
* (alignment) troubles...
*
* At least the PA7100 is rumored to be able to deal with cache-misses
* without stalling instruction issue. If this is true, and the cache is
* actually also lockup-free, we should use a deeper software pipeline, and
* load from S1 very early! (The loads and stores to -12(sp) will surely be
* in the cache.)
*/
.code
.export mpihelp_mul_1
.label mpihelp_mul_1
.proc
.callinfo frame=64,no_calls
.entry
ldo 64(%r30),%r30
fldws,ma 4(%r25),%fr5
stw %r23,-16(%r30) ; move s2_limb ...
addib,= -1,%r24,L$just_one_limb
fldws -16(%r30),%fr4 ; ... into fr4
add %r0,%r0,%r0 ; clear carry
xmpyu %fr4,%fr5,%fr6
fldws,ma 4(%r25),%fr7
fstds %fr6,-16(%r30)
xmpyu %fr4,%fr7,%fr8
ldw -12(%r30),%r19 ; least significant limb in product
ldw -16(%r30),%r28
fstds %fr8,-16(%r30)
addib,= -1,%r24,L$end
ldw -12(%r30),%r1
; Main loop
.label L$loop
fldws,ma 4(%r25),%fr5
stws,ma %r19,4(%r26)
addc %r28,%r1,%r19
xmpyu %fr4,%fr5,%fr6
ldw -16(%r30),%r28
fstds %fr6,-16(%r30)
addib,<> -1,%r24,L$loop
ldw -12(%r30),%r1
.label L$end
stws,ma %r19,4(%r26)
addc %r28,%r1,%r19
ldw -16(%r30),%r28
stws,ma %r19,4(%r26)
addc %r0,%r28,%r28
bv 0(%r2)
ldo -64(%r30),%r30
.label L$just_one_limb
xmpyu %fr4,%fr5,%fr6
fstds %fr6,-16(%r30)
ldw -16(%r30),%r28
ldo -64(%r30),%r30
bv 0(%r2)
fstws %fr6R,0(%r26)
.exit
.procend

117
mpi/hppa1.1/mpih-mul2.S Normal file
View File

@ -0,0 +1,117 @@
/* hppa1.1 addmul_1 -- Multiply a limb vector with a limb and add
* the result to a second limb vector.
*
* Copyright (C) 1992, 1993, 1994, 1998,
* 2001 Free Software Foundation, Inc.
*
* This file is part of GnuPG.
*
* GnuPG is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* GnuPG is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA
*
* Note: This code is heavily based on the GNU MP Library.
* Actually it's the same code with only minor changes in the
* way the data is stored; this is to support the abstraction
* of an optional secure memory allocation which may be used
* to avoid revealing of sensitive data due to paging etc.
* The GNU MP Library itself is published under the LGPL;
* however I decided to publish this code under the plain GPL.
*/
/*******************
* mpi_limb_t
* mpihelp_addmul_1( mpi_ptr_t res_ptr, (r26)
* mpi_ptr_t s1_ptr, (r25)
* mpi_size_t s1_size, (r24)
* mpi_limb_t s2_limb) (r23)
*
* This runs at 11 cycles/limb on a PA7000. With the used instructions, it
* can not become faster due to data cache contention after a store. On the
* PA7100 it runs at 10 cycles/limb, and that can not be improved either,
* since only the xmpyu does not need the integer pipeline, so the only
* dual-issue we will get are addc+xmpyu. Unrolling could gain a cycle/limb
* on the PA7100.
*
* There are some ideas described in mul1.S that applies to this code too.
*/
.code
.export mpihelp_addmul_1
.label mpihelp_addmul_1
.proc
.callinfo frame=64,no_calls
.entry
ldo 64(%r30),%r30
fldws,ma 4(%r25),%fr5
stw %r23,-16(%r30) ; move s2_limb ...
addib,= -1,%r24,L$just_one_limb
fldws -16(%r30),%fr4 ; ... into fr4
add %r0,%r0,%r0 ; clear carry
xmpyu %fr4,%fr5,%fr6
fldws,ma 4(%r25),%fr7
fstds %fr6,-16(%r30)
xmpyu %fr4,%fr7,%fr8
ldw -12(%r30),%r19 ; least significant limb in product
ldw -16(%r30),%r28
fstds %fr8,-16(%r30)
addib,= -1,%r24,L$end
ldw -12(%r30),%r1
; Main loop
.label L$loop
ldws 0(%r26),%r29
fldws,ma 4(%r25),%fr5
add %r29,%r19,%r19
stws,ma %r19,4(%r26)
addc %r28,%r1,%r19
xmpyu %fr4,%fr5,%fr6
ldw -16(%r30),%r28
fstds %fr6,-16(%r30)
addc %r0,%r28,%r28
addib,<> -1,%r24,L$loop
ldw -12(%r30),%r1
.label L$end
ldw 0(%r26),%r29
add %r29,%r19,%r19
stws,ma %r19,4(%r26)
addc %r28,%r1,%r19
ldw -16(%r30),%r28
ldws 0(%r26),%r29
addc %r0,%r28,%r28
add %r29,%r19,%r19
stws,ma %r19,4(%r26)
addc %r0,%r28,%r28
bv 0(%r2)
ldo -64(%r30),%r30
.label L$just_one_limb
xmpyu %fr4,%fr5,%fr6
ldw 0(%r26),%r29
fstds %fr6,-16(%r30)
ldw -12(%r30),%r1
ldw -16(%r30),%r28
add %r29,%r1,%r19
stw %r19,0(%r26)
addc %r0,%r28,%r28
bv 0(%r2)
ldo -64(%r30),%r30
.exit
.procend

127
mpi/hppa1.1/mpih-mul3.S Normal file
View File

@ -0,0 +1,127 @@
/* hppa1.1 submul_1 -- Multiply a limb vector with a limb and add
* the result to a second limb vector.
*
* Copyright (C) 1992, 1993, 1994, 1998,
* 2001 Free Software Foundation, Inc.
*
* This file is part of GnuPG.
*
* GnuPG is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* GnuPG is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA
*
* Note: This code is heavily based on the GNU MP Library.
* Actually it's the same code with only minor changes in the
* way the data is stored; this is to support the abstraction
* of an optional secure memory allocation which may be used
* to avoid revealing of sensitive data due to paging etc.
* The GNU MP Library itself is published under the LGPL;
* however I decided to publish this code under the plain GPL.
*/
/*******************
* mpi_limb_t
* mpihelp_submul_1( mpi_ptr_t res_ptr, (r26)
* mpi_ptr_t s1_ptr, (r25)
* mpi_size_t s1_size, (r24)
* mpi_limb_t s2_limb) (r23)
*
*
* This runs at 12 cycles/limb on a PA7000. With the used instructions, it
* can not become faster due to data cache contention after a store. On the
* PA7100 it runs at 11 cycles/limb, and that can not be improved either,
* since only the xmpyu does not need the integer pipeline, so the only
* dual-issue we will get are addc+xmpyu. Unrolling could gain a cycle/limb
* on the PA7100.
*
* There are some ideas described in mul1.S that applies to this code too.
*
* It seems possible to make this run as fast as addmul_1, if we use
* sub,>>= %r29,%r19,%r22
* addi 1,%r28,%r28
* but that requires reworking the hairy software pipeline...
*/
.code
.export mpihelp_submul_1
.label mpihelp_submul_1
.proc
.callinfo frame=64,no_calls
.entry
ldo 64(%r30),%r30
fldws,ma 4(%r25),%fr5
stw %r23,-16(%r30) ; move s2_limb ...
addib,= -1,%r24,L$just_one_limb
fldws -16(%r30),%fr4 ; ... into fr4
add %r0,%r0,%r0 ; clear carry
xmpyu %fr4,%fr5,%fr6
fldws,ma 4(%r25),%fr7
fstds %fr6,-16(%r30)
xmpyu %fr4,%fr7,%fr8
ldw -12(%r30),%r19 ; least significant limb in product
ldw -16(%r30),%r28
fstds %fr8,-16(%r30)
addib,= -1,%r24,L$end
ldw -12(%r30),%r1
; Main loop
.label L$loop
ldws 0(%r26),%r29
fldws,ma 4(%r25),%fr5
sub %r29,%r19,%r22
add %r22,%r19,%r0
stws,ma %r22,4(%r26)
addc %r28,%r1,%r19
xmpyu %fr4,%fr5,%fr6
ldw -16(%r30),%r28
fstds %fr6,-16(%r30)
addc %r0,%r28,%r28
addib,<> -1,%r24,L$loop
ldw -12(%r30),%r1
.label L$end
ldw 0(%r26),%r29
sub %r29,%r19,%r22
add %r22,%r19,%r0
stws,ma %r22,4(%r26)
addc %r28,%r1,%r19
ldw -16(%r30),%r28
ldws 0(%r26),%r29
addc %r0,%r28,%r28
sub %r29,%r19,%r22
add %r22,%r19,%r0
stws,ma %r22,4(%r26)
addc %r0,%r28,%r28
bv 0(%r2)
ldo -64(%r30),%r30
.label L$just_one_limb
xmpyu %fr4,%fr5,%fr6
ldw 0(%r26),%r29
fstds %fr6,-16(%r30)
ldw -12(%r30),%r1
ldw -16(%r30),%r28
sub %r29,%r1,%r22
add %r22,%r1,%r0
stw %r22,0(%r26)
addc %r0,%r28,%r28
bv 0(%r2)
ldo -64(%r30),%r30
.exit
.procend

90
mpi/hppa1.1/udiv-qrnnd.S Normal file
View File

@ -0,0 +1,90 @@
/* HP-PA __udiv_qrnnd division support, used from longlong.h.
* This version runs fast on PA 7000 and later.
*
* Copyright (C) 1993, 1994, 1998,
* 2001 Free Software Foundation, Inc.
*
* This file is part of GnuPG.
*
* GnuPG is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* GnuPG is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA
*
* Note: This code is heavily based on the GNU MP Library.
* Actually it's the same code with only minor changes in the
* way the data is stored; this is to support the abstraction
* of an optional secure memory allocation which may be used
* to avoid revealing of sensitive data due to paging etc.
* The GNU MP Library itself is published under the LGPL;
* however I decided to publish this code under the plain GPL.
*/
/* INPUT PARAMETERS
* rem_ptr gr26
* n1 gr25
* n0 gr24
* d gr23
*/
.code
.label L$0000
.word 0x43f00000
.word 0x0
.export __udiv_qrnnd
.label __udiv_qrnnd
.proc
.callinfo frame=64,no_calls
.entry
ldo 64(%r30),%r30
stws %r25,-16(0,%r30) ; n_hi
stws %r24,-12(0,%r30) ; n_lo
ldil L'L$0000,%r19 ; '
ldo R'L$0000(%r19),%r19 ; '
fldds -16(0,%r30),%fr5
stws %r23,-12(0,%r30)
comib,<= 0,%r25,L$1
fcnvxf,dbl,dbl %fr5,%fr5
fldds 0(0,%r19),%fr4
fadd,dbl %fr4,%fr5,%fr5
.label L$1
fcpy,sgl %fr0,%fr6L
fldws -12(0,%r30),%fr6R
fcnvxf,dbl,dbl %fr6,%fr4
fdiv,dbl %fr5,%fr4,%fr5
fcnvfx,dbl,dbl %fr5,%fr4
fstws %fr4R,-16(%r30)
xmpyu %fr4R,%fr6R,%fr6
ldws -16(%r30),%r28
fstds %fr6,-16(0,%r30)
ldws -12(0,%r30),%r21
ldws -16(0,%r30),%r20
sub %r24,%r21,%r22
subb %r25,%r20,%r19
comib,= 0,%r19,L$2
ldo -64(%r30),%r30
add %r22,%r23,%r22
ldo -1(%r28),%r28
.label L$2
bv 0(%r2)
stws %r22,0(0,%r26)
.exit
.procend

9
mpi/i386/distfiles Normal file
View File

@ -0,0 +1,9 @@
mpih-add1.S
mpih-mul1.S
mpih-mul2.S
mpih-mul3.S
mpih-lshift.S
mpih-rshift.S
mpih-sub1.S
syntax.h

118
mpi/i386/mpih-add1.S Normal file
View File

@ -0,0 +1,118 @@
/* i80386 add_n -- Add two limb vectors of the same length > 0 and store
* sum in a third limb vector.
*
* Copyright (C) 1992, 1994, 1995, 1998,
* 2001, 2002 Free Software Foundation, Inc.
*
* This file is part of GnuPG.
*
* GnuPG is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* GnuPG is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA
*
* Note: This code is heavily based on the GNU MP Library.
* Actually it's the same code with only minor changes in the
* way the data is stored; this is to support the abstraction
* of an optional secure memory allocation which may be used
* to avoid revealing of sensitive data due to paging etc.
* The GNU MP Library itself is published under the LGPL;
* however I decided to publish this code under the plain GPL.
*/
#include "sysdep.h"
#include "asm-syntax.h"
/*******************
* mpi_limb_t
* mpihelp_add_n( mpi_ptr_t res_ptr, (sp + 4)
* mpi_ptr_t s1_ptr, (sp + 8)
* mpi_ptr_t s2_ptr, (sp + 12)
* mpi_size_t size) (sp + 16)
*/
.text
ALIGN (3)
.globl C_SYMBOL_NAME(mpihelp_add_n)
C_SYMBOL_NAME(mpihelp_add_n:)
pushl %edi
pushl %esi
movl 12(%esp),%edi /* res_ptr */
movl 16(%esp),%esi /* s1_ptr */
movl 20(%esp),%edx /* s2_ptr */
movl 24(%esp),%ecx /* size */
movl %ecx,%eax
shrl $3,%ecx /* compute count for unrolled loop */
negl %eax
andl $7,%eax /* get index where to start loop */
jz Loop /* necessary special case for 0 */
incl %ecx /* adjust loop count */
shll $2,%eax /* adjustment for pointers... */
subl %eax,%edi /* ... since they are offset ... */
subl %eax,%esi /* ... by a constant when we ... */
subl %eax,%edx /* ... enter the loop */
shrl $2,%eax /* restore previous value */
#ifdef PIC
/* Calculate start address in loop for PIC. Due to limitations in some
assemblers, Loop-L0-3 cannot be put into the leal */
call L0
L0: leal (%eax,%eax,8),%eax
addl (%esp),%eax
addl $(Loop-L0-3),%eax
addl $4,%esp
#else
/* Calculate start address in loop for non-PIC. */
leal Loop-3(%eax,%eax,8),%eax
#endif
jmp *%eax /* jump into loop */
ALIGN (3)
Loop: movl (%esi),%eax
adcl (%edx),%eax
movl %eax,(%edi)
movl 4(%esi),%eax
adcl 4(%edx),%eax
movl %eax,4(%edi)
movl 8(%esi),%eax
adcl 8(%edx),%eax
movl %eax,8(%edi)
movl 12(%esi),%eax
adcl 12(%edx),%eax
movl %eax,12(%edi)
movl 16(%esi),%eax
adcl 16(%edx),%eax
movl %eax,16(%edi)
movl 20(%esi),%eax
adcl 20(%edx),%eax
movl %eax,20(%edi)
movl 24(%esi),%eax
adcl 24(%edx),%eax
movl %eax,24(%edi)
movl 28(%esi),%eax
adcl 28(%edx),%eax
movl %eax,28(%edi)
leal 32(%edi),%edi
leal 32(%esi),%esi
leal 32(%edx),%edx
decl %ecx
jnz Loop
sbbl %eax,%eax
negl %eax
popl %esi
popl %edi
ret

96
mpi/i386/mpih-lshift.S Normal file
View File

@ -0,0 +1,96 @@
/* i80386 lshift
* Copyright (C) 1992, 1994, 1998,
* 2001 Free Software Foundation, Inc.
*
* This file is part of GnuPG.
*
* GnuPG is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* GnuPG is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA
*
* Note: This code is heavily based on the GNU MP Library.
* Actually it's the same code with only minor changes in the
* way the data is stored; this is to support the abstraction
* of an optional secure memory allocation which may be used
* to avoid revealing of sensitive data due to paging etc.
* The GNU MP Library itself is published under the LGPL;
* however I decided to publish this code under the plain GPL.
*/
#include "sysdep.h"
#include "asm-syntax.h"
/*******************
* mpi_limb_t
* mpihelp_lshift( mpi_ptr_t wp, (sp + 4)
* mpi_ptr_t up, (sp + 8)
* mpi_size_t usize, (sp + 12)
* unsigned cnt) (sp + 16)
*/
.text
ALIGN (3)
.globl C_SYMBOL_NAME(mpihelp_lshift)
C_SYMBOL_NAME(mpihelp_lshift:)
pushl %edi
pushl %esi
pushl %ebx
movl 16(%esp),%edi /* res_ptr */
movl 20(%esp),%esi /* s_ptr */
movl 24(%esp),%edx /* size */
movl 28(%esp),%ecx /* cnt */
subl $4,%esi /* adjust s_ptr */
movl (%esi,%edx,4),%ebx /* read most significant limb */
xorl %eax,%eax
shldl %ebx,%eax /* compute carry limb */
decl %edx
jz Lend
pushl %eax /* push carry limb onto stack */
testb $1,%dl
jnz L1 /* enter loop in the middle */
movl %ebx,%eax
ALIGN (3)
Loop: movl (%esi,%edx,4),%ebx /* load next lower limb */
shldl %ebx,%eax /* compute result limb */
movl %eax,(%edi,%edx,4) /* store it */
decl %edx
L1: movl (%esi,%edx,4),%eax
shldl %eax,%ebx
movl %ebx,(%edi,%edx,4)
decl %edx
jnz Loop
shll %cl,%eax /* compute least significant limb */
movl %eax,(%edi) /* store it */
popl %eax /* pop carry limb */
popl %ebx
popl %esi
popl %edi
ret
Lend: shll %cl,%ebx /* compute least significant limb */
movl %ebx,(%edi) /* store it */
popl %ebx
popl %esi
popl %edi
ret

86
mpi/i386/mpih-mul1.S Normal file
View File

@ -0,0 +1,86 @@
/* i80386 mul_1 -- Multiply a limb vector with a limb and store
* the result in a second limb vector.
* Copyright (C) 1992, 1994, 1998,
* 2001 Free Software Foundation, Inc.
*
* This file is part of GnuPG.
*
* GnuPG is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* GnuPG is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA
*
* Note: This code is heavily based on the GNU MP Library.
* Actually it's the same code with only minor changes in the
* way the data is stored; this is to support the abstraction
* of an optional secure memory allocation which may be used
* to avoid revealing of sensitive data due to paging etc.
* The GNU MP Library itself is published under the LGPL;
* however I decided to publish this code under the plain GPL.
*/
#include "sysdep.h"
#include "asm-syntax.h"
/*******************
* mpi_limb_t
* mpihelp_mul_1( mpi_ptr_t res_ptr, (sp + 4)
* mpi_ptr_t s1_ptr, (sp + 8)
* mpi_size_t s1_size, (sp + 12)
* mpi_limb_t s2_limb) (sp + 16)
*/
#define res_ptr edi
#define s1_ptr esi
#define size ecx
#define s2_limb ebp
TEXT
ALIGN (3)
GLOBL C_SYMBOL_NAME(mpihelp_mul_1)
C_SYMBOL_NAME(mpihelp_mul_1:)
INSN1(push,l ,R(edi))
INSN1(push,l ,R(esi))
INSN1(push,l ,R(ebx))
INSN1(push,l ,R(ebp))
INSN2(mov,l ,R(res_ptr),MEM_DISP(esp,20))
INSN2(mov,l ,R(s1_ptr),MEM_DISP(esp,24))
INSN2(mov,l ,R(size),MEM_DISP(esp,28))
INSN2(mov,l ,R(s2_limb),MEM_DISP(esp,32))
INSN2(lea,l ,R(res_ptr),MEM_INDEX(res_ptr,size,4))
INSN2(lea,l ,R(s1_ptr),MEM_INDEX(s1_ptr,size,4))
INSN1(neg,l ,R(size))
INSN2(xor,l ,R(ebx),R(ebx))
ALIGN (3)
Loop:
INSN2(mov,l ,R(eax),MEM_INDEX(s1_ptr,size,4))
INSN1(mul,l ,R(s2_limb))
INSN2(add,l ,R(eax),R(ebx))
INSN2(mov,l ,MEM_INDEX(res_ptr,size,4),R(eax))
INSN2(adc,l ,R(edx),$0)
INSN2(mov,l ,R(ebx),R(edx))
INSN1(inc,l ,R(size))
INSN1(jnz, ,Loop)
INSN2(mov,l ,R(eax),R(ebx))
INSN1(pop,l ,R(ebp))
INSN1(pop,l ,R(ebx))
INSN1(pop,l ,R(esi))
INSN1(pop,l ,R(edi))
ret

88
mpi/i386/mpih-mul2.S Normal file
View File

@ -0,0 +1,88 @@
/* i80386 addmul_1 -- Multiply a limb vector with a limb and add
* the result to a second limb vector.
*
* Copyright (C) 1992, 1994, 1998,
* 2001 Free Software Foundation, Inc.
*
* This file is part of GnuPG.
*
* GnuPG is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* GnuPG is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA
*
* Note: This code is heavily based on the GNU MP Library.
* Actually it's the same code with only minor changes in the
* way the data is stored; this is to support the abstraction
* of an optional secure memory allocation which may be used
* to avoid revealing of sensitive data due to paging etc.
* The GNU MP Library itself is published under the LGPL;
* however I decided to publish this code under the plain GPL.
*/
#include "sysdep.h"
#include "asm-syntax.h"
/*******************
* mpi_limb_t
* mpihelp_addmul_1( mpi_ptr_t res_ptr, (sp + 4)
* mpi_ptr_t s1_ptr, (sp + 8)
* mpi_size_t s1_size, (sp + 12)
* mpi_limb_t s2_limb) (sp + 16)
*/
#define res_ptr edi
#define s1_ptr esi
#define size ecx
#define s2_limb ebp
TEXT
ALIGN (3)
GLOBL C_SYMBOL_NAME(mpihelp_addmul_1)
C_SYMBOL_NAME(mpihelp_addmul_1:)
INSN1(push,l ,R(edi))
INSN1(push,l ,R(esi))
INSN1(push,l ,R(ebx))
INSN1(push,l ,R(ebp))
INSN2(mov,l ,R(res_ptr),MEM_DISP(esp,20))
INSN2(mov,l ,R(s1_ptr),MEM_DISP(esp,24))
INSN2(mov,l ,R(size),MEM_DISP(esp,28))
INSN2(mov,l ,R(s2_limb),MEM_DISP(esp,32))
INSN2(lea,l ,R(res_ptr),MEM_INDEX(res_ptr,size,4))
INSN2(lea,l ,R(s1_ptr),MEM_INDEX(s1_ptr,size,4))
INSN1(neg,l ,R(size))
INSN2(xor,l ,R(ebx),R(ebx))
ALIGN (3)
Loop:
INSN2(mov,l ,R(eax),MEM_INDEX(s1_ptr,size,4))
INSN1(mul,l ,R(s2_limb))
INSN2(add,l ,R(eax),R(ebx))
INSN2(adc,l ,R(edx),$0)
INSN2(add,l ,MEM_INDEX(res_ptr,size,4),R(eax))
INSN2(adc,l ,R(edx),$0)
INSN2(mov,l ,R(ebx),R(edx))
INSN1(inc,l ,R(size))
INSN1(jnz, ,Loop)
INSN2(mov,l ,R(eax),R(ebx))
INSN1(pop,l ,R(ebp))
INSN1(pop,l ,R(ebx))
INSN1(pop,l ,R(esi))
INSN1(pop,l ,R(edi))
ret

88
mpi/i386/mpih-mul3.S Normal file
View File

@ -0,0 +1,88 @@
/* i80386 submul_1 -- Multiply a limb vector with a limb and add
* the result to a second limb vector.
*
* Copyright (C) 1992, 1994, 1998,
* 2001 Free Software Foundation, Inc.
*
* This file is part of GnuPG.
*
* GnuPG is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* GnuPG is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA
*
* Note: This code is heavily based on the GNU MP Library.
* Actually it's the same code with only minor changes in the
* way the data is stored; this is to support the abstraction
* of an optional secure memory allocation which may be used
* to avoid revealing of sensitive data due to paging etc.
* The GNU MP Library itself is published under the LGPL;
* however I decided to publish this code under the plain GPL.
*/
#include "sysdep.h"
#include "asm-syntax.h"
/*******************
* mpi_limb_t
* mpihelp_submul_1( mpi_ptr_t res_ptr, (sp + 4)
* mpi_ptr_t s1_ptr, (sp + 8)
* mpi_size_t s1_size, (sp + 12)
* mpi_limb_t s2_limb) (sp + 16)
*/
#define res_ptr edi
#define s1_ptr esi
#define size ecx
#define s2_limb ebp
TEXT
ALIGN (3)
GLOBL C_SYMBOL_NAME(mpihelp_submul_1)
C_SYMBOL_NAME(mpihelp_submul_1:)
INSN1(push,l ,R(edi))
INSN1(push,l ,R(esi))
INSN1(push,l ,R(ebx))
INSN1(push,l ,R(ebp))
INSN2(mov,l ,R(res_ptr),MEM_DISP(esp,20))
INSN2(mov,l ,R(s1_ptr),MEM_DISP(esp,24))
INSN2(mov,l ,R(size),MEM_DISP(esp,28))
INSN2(mov,l ,R(s2_limb),MEM_DISP(esp,32))
INSN2(lea,l ,R(res_ptr),MEM_INDEX(res_ptr,size,4))
INSN2(lea,l ,R(s1_ptr),MEM_INDEX(s1_ptr,size,4))
INSN1(neg,l ,R(size))
INSN2(xor,l ,R(ebx),R(ebx))
ALIGN (3)
Loop:
INSN2(mov,l ,R(eax),MEM_INDEX(s1_ptr,size,4))
INSN1(mul,l ,R(s2_limb))
INSN2(add,l ,R(eax),R(ebx))
INSN2(adc,l ,R(edx),$0)
INSN2(sub,l ,MEM_INDEX(res_ptr,size,4),R(eax))
INSN2(adc,l ,R(edx),$0)
INSN2(mov,l ,R(ebx),R(edx))
INSN1(inc,l ,R(size))
INSN1(jnz, ,Loop)
INSN2(mov,l ,R(eax),R(ebx))
INSN1(pop,l ,R(ebp))
INSN1(pop,l ,R(ebx))
INSN1(pop,l ,R(esi))
INSN1(pop,l ,R(edi))
ret

99
mpi/i386/mpih-rshift.S Normal file
View File

@ -0,0 +1,99 @@
/* i80386 rshift
*
* Copyright (C) 1992, 1994, 1998,
* 2001, 2002 Free Software Foundation, Inc.
*
* This file is part of GnuPG.
*
* GnuPG is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* GnuPG is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA
*
* Note: This code is heavily based on the GNU MP Library.
* Actually it's the same code with only minor changes in the
* way the data is stored; this is to support the abstraction
* of an optional secure memory allocation which may be used
* to avoid revealing of sensitive data due to paging etc.
* The GNU MP Library itself is published under the LGPL;
* however I decided to publish this code under the plain GPL.
*/
#include "sysdep.h"
#include "asm-syntax.h"
/*******************
* mpi_limb_t
* mpihelp_rshift( mpi_ptr_t wp, (sp + 4)
* mpi_ptr_t up, (sp + 8)
* mpi_size_t usize, (sp + 12)
* unsigned cnt) (sp + 16)
*/
.text
ALIGN (3)
.globl C_SYMBOL_NAME(mpihelp_rshift)
C_SYMBOL_NAME(mpihelp_rshift:)
pushl %edi
pushl %esi
pushl %ebx
movl 16(%esp),%edi /* wp */
movl 20(%esp),%esi /* up */
movl 24(%esp),%edx /* usize */
movl 28(%esp),%ecx /* cnt */
leal -4(%edi,%edx,4),%edi
leal (%esi,%edx,4),%esi
negl %edx
movl (%esi,%edx,4),%ebx /* read least significant limb */
xorl %eax,%eax
shrdl %ebx,%eax /* compute carry limb */
incl %edx
jz Lend2
pushl %eax /* push carry limb onto stack */
testb $1,%dl
jnz L2 /* enter loop in the middle */
movl %ebx,%eax
ALIGN (3)
Loop2: movl (%esi,%edx,4),%ebx /* load next higher limb */
shrdl %ebx,%eax /* compute result limb */
movl %eax,(%edi,%edx,4) /* store it */
incl %edx
L2: movl (%esi,%edx,4),%eax
shrdl %eax,%ebx
movl %ebx,(%edi,%edx,4)
incl %edx
jnz Loop2
shrl %cl,%eax /* compute most significant limb */
movl %eax,(%edi) /* store it */
popl %eax /* pop carry limb */
popl %ebx
popl %esi
popl %edi
ret
Lend2: shrl %cl,%ebx /* compute most significant limb */
movl %ebx,(%edi) /* store it */
popl %ebx
popl %esi
popl %edi
ret

119
mpi/i386/mpih-sub1.S Normal file
View File

@ -0,0 +1,119 @@
/* i80386 sub_n -- Sub two limb vectors of the same length > 0 and store
* sum in a third limb vector.
*
* Copyright (C) 1992, 1994, 1995, 1998,
* 2001, 2002 Free Software Foundation, Inc.
*
* This file is part of GnuPG.
*
* GnuPG is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* GnuPG is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA
*
* Note: This code is heavily based on the GNU MP Library.
* Actually it's the same code with only minor changes in the
* way the data is stored; this is to support the abstraction
* of an optional secure memory allocation which may be used
* to avoid revealing of sensitive data due to paging etc.
* The GNU MP Library itself is published under the LGPL;
* however I decided to publish this code under the plain GPL.
*/
#include "sysdep.h"
#include "asm-syntax.h"
/*******************
* mpi_limb_t
* mpihelp_sub_n( mpi_ptr_t res_ptr, (sp + 4)
* mpi_ptr_t s1_ptr, (sp + 8)
* mpi_ptr_t s2_ptr, (sp + 12)
* mpi_size_t size) (sp + 16)
*/
.text
ALIGN (3)
.globl C_SYMBOL_NAME(mpihelp_sub_n)
C_SYMBOL_NAME(mpihelp_sub_n:)
pushl %edi
pushl %esi
movl 12(%esp),%edi /* res_ptr */
movl 16(%esp),%esi /* s1_ptr */
movl 20(%esp),%edx /* s2_ptr */
movl 24(%esp),%ecx /* size */
movl %ecx,%eax
shrl $3,%ecx /* compute count for unrolled loop */
negl %eax
andl $7,%eax /* get index where to start loop */
jz Loop /* necessary special case for 0 */
incl %ecx /* adjust loop count */
shll $2,%eax /* adjustment for pointers... */
subl %eax,%edi /* ... since they are offset ... */
subl %eax,%esi /* ... by a constant when we ... */
subl %eax,%edx /* ... enter the loop */
shrl $2,%eax /* restore previous value */
#ifdef PIC
/* Calculate start address in loop for PIC. Due to limitations in some
assemblers, Loop-L0-3 cannot be put into the leal */
call L0
L0: leal (%eax,%eax,8),%eax
addl (%esp),%eax
addl $(Loop-L0-3),%eax
addl $4,%esp
#else
/* Calculate start address in loop for non-PIC. */
leal Loop-3(%eax,%eax,8),%eax
#endif
jmp *%eax /* jump into loop */
ALIGN (3)
Loop: movl (%esi),%eax
sbbl (%edx),%eax
movl %eax,(%edi)
movl 4(%esi),%eax
sbbl 4(%edx),%eax
movl %eax,4(%edi)
movl 8(%esi),%eax
sbbl 8(%edx),%eax
movl %eax,8(%edi)
movl 12(%esi),%eax
sbbl 12(%edx),%eax
movl %eax,12(%edi)
movl 16(%esi),%eax
sbbl 16(%edx),%eax
movl %eax,16(%edi)
movl 20(%esi),%eax
sbbl 20(%edx),%eax
movl %eax,20(%edi)
movl 24(%esi),%eax
sbbl 24(%edx),%eax
movl %eax,24(%edi)
movl 28(%esi),%eax
sbbl 28(%edx),%eax
movl %eax,28(%edi)
leal 32(%edi),%edi
leal 32(%esi),%esi
leal 32(%edx),%edx
decl %ecx
jnz Loop
sbbl %eax,%eax
negl %eax
popl %esi
popl %edi
ret

68
mpi/i386/syntax.h Normal file
View File

@ -0,0 +1,68 @@
/* syntax.h -- Definitions for x86 syntax variations.
*
* Copyright (C) 1992, 1994, 1995, 1998,
* 2001 Free Software Foundation, Inc.
*
* GnuPG is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* GnuPG is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA
*
* Note: This code is heavily based on the GNU MP Library.
* Actually it's the same code with only minor changes in the
* way the data is stored; this is to support the abstraction
* of an optional secure memory allocation which may be used
* to avoid revealing of sensitive data due to paging etc.
* The GNU MP Library itself is published under the LGPL;
* however I decided to publish this code under the plain GPL.
*/
#undef ALIGN
#if defined (BSD_SYNTAX) || defined (ELF_SYNTAX)
#define R(r) %r
#define MEM(base)(base)
#define MEM_DISP(base,displacement)displacement(R(base))
#define MEM_INDEX(base,index,size)(R(base),R(index),size)
#ifdef __STDC__
#define INSN1(mnemonic,size_suffix,dst)mnemonic##size_suffix dst
#define INSN2(mnemonic,size_suffix,dst,src)mnemonic##size_suffix src,dst
#else
#define INSN1(mnemonic,size_suffix,dst)mnemonic/**/size_suffix dst
#define INSN2(mnemonic,size_suffix,dst,src)mnemonic/**/size_suffix src,dst
#endif
#define TEXT .text
#if defined (BSD_SYNTAX)
#define ALIGN(log) .align log
#endif
#if defined (ELF_SYNTAX)
#define ALIGN(log) .align 1<<log
#endif
#define GLOBL .globl
#endif
#ifdef INTEL_SYNTAX
#define R(r) r
#define MEM(base)[base]
#define MEM_DISP(base,displacement)[base+(displacement)]
#define MEM_INDEX(base,index,size)[base+index*size]
#define INSN1(mnemonic,size_suffix,dst)mnemonic dst
#define INSN2(mnemonic,size_suffix,dst,src)mnemonic dst,src
#define TEXT .text
#define ALIGN(log) .align log
#define GLOBL .globl
#endif
#ifdef X86_BROKEN_ALIGN
#undef ALIGN
#define ALIGN(log) .align log,0x90
#endif

26
mpi/i586/README Normal file
View File

@ -0,0 +1,26 @@
This directory contains mpn functions optimized for Intel Pentium
processors.
RELEVANT OPTIMIZATION ISSUES
1. Pentium doesn't allocate cache lines on writes, unlike most other modern
processors. Since the functions in the mpn class do array writes, we have to
handle allocating the destination cache lines by reading a word from it in the
loops, to achieve the best performance.
2. Pairing of memory operations requires that the two issued operations refer
to different cache banks. The simplest way to insure this is to read/write
two words from the same object. If we make operations on different objects,
they might or might not be to the same cache bank.
STATUS
1. mpn_lshift and mpn_rshift run at about 6 cycles/limb, but the Pentium
documentation indicates that they should take only 43/8 = 5.375 cycles/limb,
or 5 cycles/limb asymptotically.
2. mpn_add_n and mpn_sub_n run at asymptotically 2 cycles/limb. Due to loop
overhead and other delays (cache refill?), they run at or near 2.5 cycles/limb.
3. mpn_mul_1, mpn_addmul_1, mpn_submul_1 all run 1 cycle faster than they
should...

9
mpi/i586/distfiles Normal file
View File

@ -0,0 +1,9 @@
mpih-add1.S
mpih-mul1.S
mpih-mul2.S
mpih-mul3.S
mpih-lshift.S
mpih-rshift.S
mpih-sub1.S
README

135
mpi/i586/mpih-add1.S Normal file
View File

@ -0,0 +1,135 @@
/* i80586 add_n -- Add two limb vectors of the same length > 0 and store
* sum in a third limb vector.
*
* Copyright (C) 1992, 1994, 1995, 1996, 1998,
* 2001 Free Software Foundation, Inc.
*
* This file is part of GnuPG.
*
* GnuPG is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* GnuPG is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA
*/
#include "sysdep.h"
#include "asm-syntax.h"
/*******************
* mpi_limb_t
* mpihelp_add_n( mpi_ptr_t res_ptr, (sp + 4)
* mpi_ptr_t s1_ptr, (sp + 8)
* mpi_ptr_t s2_ptr, (sp + 12)
* mpi_size_t size) (sp + 16)
*/
.text
ALIGN (3)
.globl C_SYMBOL_NAME(mpihelp_add_n)
C_SYMBOL_NAME(mpihelp_add_n:)
pushl %edi
pushl %esi
pushl %ebx
pushl %ebp
movl 20(%esp),%edi /* res_ptr */
movl 24(%esp),%esi /* s1_ptr */
movl 28(%esp),%ebp /* s2_ptr */
movl 32(%esp),%ecx /* size */
movl (%ebp),%ebx
decl %ecx
movl %ecx,%edx
shrl $3,%ecx
andl $7,%edx
testl %ecx,%ecx /* zero carry flag */
jz Lend
pushl %edx
ALIGN (3)
Loop: movl 28(%edi),%eax /* fetch destination cache line */
leal 32(%edi),%edi
L1: movl (%esi),%eax
movl 4(%esi),%edx
adcl %ebx,%eax
movl 4(%ebp),%ebx
adcl %ebx,%edx
movl 8(%ebp),%ebx
movl %eax,-32(%edi)
movl %edx,-28(%edi)
L2: movl 8(%esi),%eax
movl 12(%esi),%edx
adcl %ebx,%eax
movl 12(%ebp),%ebx
adcl %ebx,%edx
movl 16(%ebp),%ebx
movl %eax,-24(%edi)
movl %edx,-20(%edi)
L3: movl 16(%esi),%eax
movl 20(%esi),%edx
adcl %ebx,%eax
movl 20(%ebp),%ebx
adcl %ebx,%edx
movl 24(%ebp),%ebx
movl %eax,-16(%edi)
movl %edx,-12(%edi)
L4: movl 24(%esi),%eax
movl 28(%esi),%edx
adcl %ebx,%eax
movl 28(%ebp),%ebx
adcl %ebx,%edx
movl 32(%ebp),%ebx
movl %eax,-8(%edi)
movl %edx,-4(%edi)
leal 32(%esi),%esi
leal 32(%ebp),%ebp
decl %ecx
jnz Loop
popl %edx
Lend:
decl %edx /* test %edx w/o clobbering carry */
js Lend2
incl %edx
Loop2:
leal 4(%edi),%edi
movl (%esi),%eax
adcl %ebx,%eax
movl 4(%ebp),%ebx
movl %eax,-4(%edi)
leal 4(%esi),%esi
leal 4(%ebp),%ebp
decl %edx
jnz Loop2
Lend2:
movl (%esi),%eax
adcl %ebx,%eax
movl %eax,(%edi)
sbbl %eax,%eax
negl %eax
popl %ebp
popl %ebx
popl %esi
popl %edi
ret

231
mpi/i586/mpih-lshift.S Normal file
View File

@ -0,0 +1,231 @@
/* i80586 lshift
*
* Copyright (C) 1992, 1994, 1998,
* 2001 Free Software Foundation, Inc.
*
* This file is part of GnuPG.
*
* GnuPG is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* GnuPG is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA
*
* Note: This code is heavily based on the GNU MP Library.
* Actually it's the same code with only minor changes in the
* way the data is stored; this is to support the abstraction
* of an optional secure memory allocation which may be used
* to avoid revealing of sensitive data due to paging etc.
* The GNU MP Library itself is published under the LGPL;
* however I decided to publish this code under the plain GPL.
*/
#include "sysdep.h"
#include "asm-syntax.h"
/*******************
* mpi_limb_t
* mpihelp_lshift( mpi_ptr_t wp, (sp + 4)
* mpi_ptr_t up, (sp + 8)
* mpi_size_t usize, (sp + 12)
* unsigned cnt) (sp + 16)
*/
.text
ALIGN (3)
.globl C_SYMBOL_NAME(mpihelp_lshift)
C_SYMBOL_NAME(mpihelp_lshift:)
pushl %edi
pushl %esi
pushl %ebx
pushl %ebp
movl 20(%esp),%edi /* res_ptr */
movl 24(%esp),%esi /* s_ptr */
movl 28(%esp),%ebp /* size */
movl 32(%esp),%ecx /* cnt */
/* We can use faster code for shift-by-1 under certain conditions. */
cmp $1,%ecx
jne Lnormal
leal 4(%esi),%eax
cmpl %edi,%eax
jnc Lspecial /* jump if s_ptr + 1 >= res_ptr */
leal (%esi,%ebp,4),%eax
cmpl %eax,%edi
jnc Lspecial /* jump if res_ptr >= s_ptr + size */
Lnormal:
leal -4(%edi,%ebp,4),%edi
leal -4(%esi,%ebp,4),%esi
movl (%esi),%edx
subl $4,%esi
xorl %eax,%eax
shldl %cl,%edx,%eax /* compute carry limb */
pushl %eax /* push carry limb onto stack */
decl %ebp
pushl %ebp
shrl $3,%ebp
jz Lend
movl (%edi),%eax /* fetch destination cache line */
ALIGN (2)
Loop: movl -28(%edi),%eax /* fetch destination cache line */
movl %edx,%ebx
movl (%esi),%eax
movl -4(%esi),%edx
shldl %cl,%eax,%ebx
shldl %cl,%edx,%eax
movl %ebx,(%edi)
movl %eax,-4(%edi)
movl -8(%esi),%ebx
movl -12(%esi),%eax
shldl %cl,%ebx,%edx
shldl %cl,%eax,%ebx
movl %edx,-8(%edi)
movl %ebx,-12(%edi)
movl -16(%esi),%edx
movl -20(%esi),%ebx
shldl %cl,%edx,%eax
shldl %cl,%ebx,%edx
movl %eax,-16(%edi)
movl %edx,-20(%edi)
movl -24(%esi),%eax
movl -28(%esi),%edx
shldl %cl,%eax,%ebx
shldl %cl,%edx,%eax
movl %ebx,-24(%edi)
movl %eax,-28(%edi)
subl $32,%esi
subl $32,%edi
decl %ebp
jnz Loop
Lend: popl %ebp
andl $7,%ebp
jz Lend2
Loop2: movl (%esi),%eax
shldl %cl,%eax,%edx
movl %edx,(%edi)
movl %eax,%edx
subl $4,%esi
subl $4,%edi
decl %ebp
jnz Loop2
Lend2: shll %cl,%edx /* compute least significant limb */
movl %edx,(%edi) /* store it */
popl %eax /* pop carry limb */
popl %ebp
popl %ebx
popl %esi
popl %edi
ret
/* We loop from least significant end of the arrays, which is only
permissable if the source and destination don't overlap, since the
function is documented to work for overlapping source and destination.
*/
Lspecial:
movl (%esi),%edx
addl $4,%esi
decl %ebp
pushl %ebp
shrl $3,%ebp
addl %edx,%edx
incl %ebp
decl %ebp
jz LLend
movl (%edi),%eax /* fetch destination cache line */
ALIGN (2)
LLoop: movl 28(%edi),%eax /* fetch destination cache line */
movl %edx,%ebx
movl (%esi),%eax
movl 4(%esi),%edx
adcl %eax,%eax
movl %ebx,(%edi)
adcl %edx,%edx
movl %eax,4(%edi)
movl 8(%esi),%ebx
movl 12(%esi),%eax
adcl %ebx,%ebx
movl %edx,8(%edi)
adcl %eax,%eax
movl %ebx,12(%edi)
movl 16(%esi),%edx
movl 20(%esi),%ebx
adcl %edx,%edx
movl %eax,16(%edi)
adcl %ebx,%ebx
movl %edx,20(%edi)
movl 24(%esi),%eax
movl 28(%esi),%edx
adcl %eax,%eax
movl %ebx,24(%edi)
adcl %edx,%edx
movl %eax,28(%edi)
leal 32(%esi),%esi /* use leal not to clobber carry */
leal 32(%edi),%edi
decl %ebp
jnz LLoop
LLend: popl %ebp
sbbl %eax,%eax /* save carry in %eax */
andl $7,%ebp
jz LLend2
addl %eax,%eax /* restore carry from eax */
LLoop2: movl %edx,%ebx
movl (%esi),%edx
adcl %edx,%edx
movl %ebx,(%edi)
leal 4(%esi),%esi /* use leal not to clobber carry */
leal 4(%edi),%edi
decl %ebp
jnz LLoop2
jmp LL1
LLend2: addl %eax,%eax /* restore carry from eax */
LL1: movl %edx,(%edi) /* store last limb */
sbbl %eax,%eax
negl %eax
popl %ebp
popl %ebx
popl %esi
popl %edi
ret

91
mpi/i586/mpih-mul1.S Normal file
View File

@ -0,0 +1,91 @@
/* i80586 mul_1 -- Multiply a limb vector with a limb and store
* the result in a second limb vector.
*
* Copyright (C) 1992, 1994, 1996, 1998,
* 2001 Free Software Foundation, Inc.
*
* This file is part of GnuPG.
*
* GnuPG is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* GnuPG is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA
*
* Note: This code is heavily based on the GNU MP Library.
* Actually it's the same code with only minor changes in the
* way the data is stored; this is to support the abstraction
* of an optional secure memory allocation which may be used
* to avoid revealing of sensitive data due to paging etc.
* The GNU MP Library itself is published under the LGPL;
* however I decided to publish this code under the plain GPL.
*/
#include "sysdep.h"
#include "asm-syntax.h"
/*******************
* mpi_limb_t
* mpihelp_mul_1( mpi_ptr_t res_ptr, (sp + 4)
* mpi_ptr_t s1_ptr, (sp + 8)
* mpi_size_t s1_size, (sp + 12)
* mpi_limb_t s2_limb) (sp + 16)
*/
#define res_ptr edi
#define s1_ptr esi
#define size ecx
#define s2_limb ebp
TEXT
ALIGN (3)
GLOBL C_SYMBOL_NAME(mpihelp_mul_1)
C_SYMBOL_NAME(mpihelp_mul_1:)
INSN1(push,l ,R(edi))
INSN1(push,l ,R(esi))
INSN1(push,l ,R(ebx))
INSN1(push,l ,R(ebp))
INSN2(mov,l ,R(res_ptr),MEM_DISP(esp,20))
INSN2(mov,l ,R(s1_ptr),MEM_DISP(esp,24))
INSN2(mov,l ,R(size),MEM_DISP(esp,28))
INSN2(mov,l ,R(s2_limb),MEM_DISP(esp,32))
INSN2(lea,l ,R(res_ptr),MEM_INDEX(res_ptr,size,4))
INSN2(lea,l ,R(s1_ptr),MEM_INDEX(s1_ptr,size,4))
INSN1(neg,l ,R(size))
INSN2(xor,l ,R(ebx),R(ebx))
ALIGN (3)
Loop: INSN2(adc,l ,R(ebx),$0)
INSN2(mov,l ,R(eax),MEM_INDEX(s1_ptr,size,4))
INSN1(mul,l ,R(s2_limb))
INSN2(add,l ,R(ebx),R(eax))
INSN2(mov,l ,MEM_INDEX(res_ptr,size,4),R(ebx))
INSN1(inc,l ,R(size))
INSN2(mov,l ,R(ebx),R(edx))
INSN1(jnz, ,Loop)
INSN2(adc,l ,R(ebx),$0)
INSN2(mov,l ,R(eax),R(ebx))
INSN1(pop,l ,R(ebp))
INSN1(pop,l ,R(ebx))
INSN1(pop,l ,R(esi))
INSN1(pop,l ,R(edi))
ret

95
mpi/i586/mpih-mul2.S Normal file
View File

@ -0,0 +1,95 @@
/* i80586 addmul_1 -- Multiply a limb vector with a limb and add
* the result to a second limb vector.
*
* Copyright (C) 1992, 1994, 1998,
* 2001 Free Software Foundation, Inc.
*
* This file is part of GnuPG.
*
* GnuPG is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* GnuPG is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA
*
* Note: This code is heavily based on the GNU MP Library.
* Actually it's the same code with only minor changes in the
* way the data is stored; this is to support the abstraction
* of an optional secure memory allocation which may be used
* to avoid revealing of sensitive data due to paging etc.
* The GNU MP Library itself is published under the LGPL;
* however I decided to publish this code under the plain GPL.
*/
#include "sysdep.h"
#include "asm-syntax.h"
/*******************
* mpi_limb_t
* mpihelp_addmul_1( mpi_ptr_t res_ptr, (sp + 4)
* mpi_ptr_t s1_ptr, (sp + 8)
* mpi_size_t s1_size, (sp + 12)
* mpi_limb_t s2_limb) (sp + 16)
*/
#define res_ptr edi
#define s1_ptr esi
#define size ecx
#define s2_limb ebp
TEXT
ALIGN (3)
GLOBL C_SYMBOL_NAME(mpihelp_addmul_1)
C_SYMBOL_NAME(mpihelp_addmul_1:)
INSN1(push,l ,R(edi))
INSN1(push,l ,R(esi))
INSN1(push,l ,R(ebx))
INSN1(push,l ,R(ebp))
INSN2(mov,l ,R(res_ptr),MEM_DISP(esp,20))
INSN2(mov,l ,R(s1_ptr),MEM_DISP(esp,24))
INSN2(mov,l ,R(size),MEM_DISP(esp,28))
INSN2(mov,l ,R(s2_limb),MEM_DISP(esp,32))
INSN2(lea,l ,R(res_ptr),MEM_INDEX(res_ptr,size,4))
INSN2(lea,l ,R(s1_ptr),MEM_INDEX(s1_ptr,size,4))
INSN1(neg,l ,R(size))
INSN2(xor,l ,R(ebx),R(ebx))
ALIGN (3)
Loop: INSN2(adc,l ,R(ebx),$0)
INSN2(mov,l ,R(eax),MEM_INDEX(s1_ptr,size,4))
INSN1(mul,l ,R(s2_limb))
INSN2(add,l ,R(eax),R(ebx))
INSN2(mov,l ,R(ebx),MEM_INDEX(res_ptr,size,4))
INSN2(adc,l ,R(edx),$0)
INSN2(add,l ,R(ebx),R(eax))
INSN2(mov,l ,MEM_INDEX(res_ptr,size,4),R(ebx))
INSN1(inc,l ,R(size))
INSN2(mov,l ,R(ebx),R(edx))
INSN1(jnz, ,Loop)
INSN2(adc,l ,R(ebx),$0)
INSN2(mov,l ,R(eax),R(ebx))
INSN1(pop,l ,R(ebp))
INSN1(pop,l ,R(ebx))
INSN1(pop,l ,R(esi))
INSN1(pop,l ,R(edi))
ret

95
mpi/i586/mpih-mul3.S Normal file
View File

@ -0,0 +1,95 @@
/* i80586 submul_1 -- Multiply a limb vector with a limb and add
* the result to a second limb vector.
*
* Copyright (C) 1992, 1994, 1998,
* 2001 Free Software Foundation, Inc.
*
* This file is part of GnuPG.
*
* GnuPG is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* GnuPG is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA
*
* Note: This code is heavily based on the GNU MP Library.
* Actually it's the same code with only minor changes in the
* way the data is stored; this is to support the abstraction
* of an optional secure memory allocation which may be used
* to avoid revealing of sensitive data due to paging etc.
* The GNU MP Library itself is published under the LGPL;
* however I decided to publish this code under the plain GPL.
*/
#include "sysdep.h"
#include "asm-syntax.h"
/*******************
* mpi_limb_t
* mpihelp_submul_1( mpi_ptr_t res_ptr, (sp + 4)
* mpi_ptr_t s1_ptr, (sp + 8)
* mpi_size_t s1_size, (sp + 12)
* mpi_limb_t s2_limb) (sp + 16)
*/
#define res_ptr edi
#define s1_ptr esi
#define size ecx
#define s2_limb ebp
TEXT
ALIGN (3)
GLOBL C_SYMBOL_NAME(mpihelp_submul_1)
C_SYMBOL_NAME(mpihelp_submul_1:)
INSN1(push,l ,R(edi))
INSN1(push,l ,R(esi))
INSN1(push,l ,R(ebx))
INSN1(push,l ,R(ebp))
INSN2(mov,l ,R(res_ptr),MEM_DISP(esp,20))
INSN2(mov,l ,R(s1_ptr),MEM_DISP(esp,24))
INSN2(mov,l ,R(size),MEM_DISP(esp,28))
INSN2(mov,l ,R(s2_limb),MEM_DISP(esp,32))
INSN2(lea,l ,R(res_ptr),MEM_INDEX(res_ptr,size,4))
INSN2(lea,l ,R(s1_ptr),MEM_INDEX(s1_ptr,size,4))
INSN1(neg,l ,R(size))
INSN2(xor,l ,R(ebx),R(ebx))
ALIGN (3)
Loop: INSN2(adc,l ,R(ebx),$0)
INSN2(mov,l ,R(eax),MEM_INDEX(s1_ptr,size,4))
INSN1(mul,l ,R(s2_limb))
INSN2(add,l ,R(eax),R(ebx))
INSN2(mov,l ,R(ebx),MEM_INDEX(res_ptr,size,4))
INSN2(adc,l ,R(edx),$0)
INSN2(sub,l ,R(ebx),R(eax))
INSN2(mov,l ,MEM_INDEX(res_ptr,size,4),R(ebx))
INSN1(inc,l ,R(size))
INSN2(mov,l ,R(ebx),R(edx))
INSN1(jnz, ,Loop)
INSN2(adc,l ,R(ebx),$0)
INSN2(mov,l ,R(eax),R(ebx))
INSN1(pop,l ,R(ebp))
INSN1(pop,l ,R(ebx))
INSN1(pop,l ,R(esi))
INSN1(pop,l ,R(edi))
ret

230
mpi/i586/mpih-rshift.S Normal file
View File

@ -0,0 +1,230 @@
/* i80586 rshift
*
* Copyright (C) 1992, 1994, 1998,
* 2001 Free Software Foundation, Inc.
*
* This file is part of GnuPG.
*
* GnuPG is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* GnuPG is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA
*
* Note: This code is heavily based on the GNU MP Library.
* Actually it's the same code with only minor changes in the
* way the data is stored; this is to support the abstraction
* of an optional secure memory allocation which may be used
* to avoid revealing of sensitive data due to paging etc.
* The GNU MP Library itself is published under the LGPL;
* however I decided to publish this code under the plain GPL.
*/
#include "sysdep.h"
#include "asm-syntax.h"
/*******************
* mpi_limb_t
* mpihelp_rshift( mpi_ptr_t wp, (sp + 4)
* mpi_ptr_t up, (sp + 8)
* mpi_size_t usize, (sp + 12)
* unsigned cnt) (sp + 16)
*/
.text
ALIGN (3)
.globl C_SYMBOL_NAME(mpihelp_rshift)
C_SYMBOL_NAME(mpihelp_rshift:)
pushl %edi
pushl %esi
pushl %ebx
pushl %ebp
movl 20(%esp),%edi /* res_ptr */
movl 24(%esp),%esi /* s_ptr */
movl 28(%esp),%ebp /* size */
movl 32(%esp),%ecx /* cnt */
/* We can use faster code for shift-by-1 under certain conditions. */
cmp $1,%ecx
jne Rnormal
leal 4(%edi),%eax
cmpl %esi,%eax
jnc Rspecial /* jump if res_ptr + 1 >= s_ptr */
leal (%edi,%ebp,4),%eax
cmpl %eax,%esi
jnc Rspecial /* jump if s_ptr >= res_ptr + size */
Rnormal:
movl (%esi),%edx
addl $4,%esi
xorl %eax,%eax
shrdl %cl,%edx,%eax /* compute carry limb */
pushl %eax /* push carry limb onto stack */
decl %ebp
pushl %ebp
shrl $3,%ebp
jz Rend
movl (%edi),%eax /* fetch destination cache line */
ALIGN (2)
Roop: movl 28(%edi),%eax /* fetch destination cache line */
movl %edx,%ebx
movl (%esi),%eax
movl 4(%esi),%edx
shrdl %cl,%eax,%ebx
shrdl %cl,%edx,%eax
movl %ebx,(%edi)
movl %eax,4(%edi)
movl 8(%esi),%ebx
movl 12(%esi),%eax
shrdl %cl,%ebx,%edx
shrdl %cl,%eax,%ebx
movl %edx,8(%edi)
movl %ebx,12(%edi)
movl 16(%esi),%edx
movl 20(%esi),%ebx
shrdl %cl,%edx,%eax
shrdl %cl,%ebx,%edx
movl %eax,16(%edi)
movl %edx,20(%edi)
movl 24(%esi),%eax
movl 28(%esi),%edx
shrdl %cl,%eax,%ebx
shrdl %cl,%edx,%eax
movl %ebx,24(%edi)
movl %eax,28(%edi)
addl $32,%esi
addl $32,%edi
decl %ebp
jnz Roop
Rend: popl %ebp
andl $7,%ebp
jz Rend2
Roop2: movl (%esi),%eax
shrdl %cl,%eax,%edx /* compute result limb */
movl %edx,(%edi)
movl %eax,%edx
addl $4,%esi
addl $4,%edi
decl %ebp
jnz Roop2
Rend2: shrl %cl,%edx /* compute most significant limb */
movl %edx,(%edi) /* store it */
popl %eax /* pop carry limb */
popl %ebp
popl %ebx
popl %esi
popl %edi
ret
/* We loop from least significant end of the arrays, which is only
permissable if the source and destination don't overlap, since the
function is documented to work for overlapping source and destination.
*/
Rspecial:
leal -4(%edi,%ebp,4),%edi
leal -4(%esi,%ebp,4),%esi
movl (%esi),%edx
subl $4,%esi
decl %ebp
pushl %ebp
shrl $3,%ebp
shrl $1,%edx
incl %ebp
decl %ebp
jz RLend
movl (%edi),%eax /* fetch destination cache line */
ALIGN (2)
RLoop: movl -28(%edi),%eax /* fetch destination cache line */
movl %edx,%ebx
movl (%esi),%eax
movl -4(%esi),%edx
rcrl $1,%eax
movl %ebx,(%edi)
rcrl $1,%edx
movl %eax,-4(%edi)
movl -8(%esi),%ebx
movl -12(%esi),%eax
rcrl $1,%ebx
movl %edx,-8(%edi)
rcrl $1,%eax
movl %ebx,-12(%edi)
movl -16(%esi),%edx
movl -20(%esi),%ebx
rcrl $1,%edx
movl %eax,-16(%edi)
rcrl $1,%ebx
movl %edx,-20(%edi)
movl -24(%esi),%eax
movl -28(%esi),%edx
rcrl $1,%eax
movl %ebx,-24(%edi)
rcrl $1,%edx
movl %eax,-28(%edi)
leal -32(%esi),%esi /* use leal not to clobber carry */
leal -32(%edi),%edi
decl %ebp
jnz RLoop
RLend: popl %ebp
sbbl %eax,%eax /* save carry in %eax */
andl $7,%ebp
jz RLend2
addl %eax,%eax /* restore carry from eax */
RLoop2: movl %edx,%ebx
movl (%esi),%edx
rcrl $1,%edx
movl %ebx,(%edi)
leal -4(%esi),%esi /* use leal not to clobber carry */
leal -4(%edi),%edi
decl %ebp
jnz RLoop2
jmp RL1
RLend2: addl %eax,%eax /* restore carry from eax */
RL1: movl %edx,(%edi) /* store last limb */
movl $0,%eax
rcrl $1,%eax
popl %ebp
popl %ebx
popl %esi
popl %edi
ret

144
mpi/i586/mpih-sub1.S Normal file
View File

@ -0,0 +1,144 @@
/* i80586 sub_n -- Sub two limb vectors of the same length > 0 and store
* sum in a third limb vector.
*
* Copyright (C) 1992, 1994, 1995, 1998,
* 2001 Free Software Foundation, Inc.
*
* This file is part of GnuPG.
*
* GnuPG is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* GnuPG is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA
*
* Note: This code is heavily based on the GNU MP Library.
* Actually it's the same code with only minor changes in the
* way the data is stored; this is to support the abstraction
* of an optional secure memory allocation which may be used
* to avoid revealing of sensitive data due to paging etc.
* The GNU MP Library itself is published under the LGPL;
* however I decided to publish this code under the plain GPL.
*/
#include "sysdep.h"
#include "asm-syntax.h"
/*******************
* mpi_limb_t
* mpihelp_sub_n( mpi_ptr_t res_ptr, (sp + 4)
* mpi_ptr_t s1_ptr, (sp + 8)
* mpi_ptr_t s2_ptr, (sp + 12)
* mpi_size_t size) (sp + 16)
*/
.text
ALIGN (3)
.globl C_SYMBOL_NAME(mpihelp_sub_n)
C_SYMBOL_NAME(mpihelp_sub_n:)
pushl %edi
pushl %esi
pushl %ebx
pushl %ebp
movl 20(%esp),%edi /* res_ptr */
movl 24(%esp),%esi /* s1_ptr */
movl 28(%esp),%ebp /* s2_ptr */
movl 32(%esp),%ecx /* size */
movl (%ebp),%ebx
decl %ecx
movl %ecx,%edx
shrl $3,%ecx
andl $7,%edx
testl %ecx,%ecx /* zero carry flag */
jz Lend
pushl %edx
ALIGN (3)
Loop: movl 28(%edi),%eax /* fetch destination cache line */
leal 32(%edi),%edi
L1: movl (%esi),%eax
movl 4(%esi),%edx
sbbl %ebx,%eax
movl 4(%ebp),%ebx
sbbl %ebx,%edx
movl 8(%ebp),%ebx
movl %eax,-32(%edi)
movl %edx,-28(%edi)
L2: movl 8(%esi),%eax
movl 12(%esi),%edx
sbbl %ebx,%eax
movl 12(%ebp),%ebx
sbbl %ebx,%edx
movl 16(%ebp),%ebx
movl %eax,-24(%edi)
movl %edx,-20(%edi)
L3: movl 16(%esi),%eax
movl 20(%esi),%edx
sbbl %ebx,%eax
movl 20(%ebp),%ebx
sbbl %ebx,%edx
movl 24(%ebp),%ebx
movl %eax,-16(%edi)
movl %edx,-12(%edi)
L4: movl 24(%esi),%eax
movl 28(%esi),%edx
sbbl %ebx,%eax
movl 28(%ebp),%ebx
sbbl %ebx,%edx
movl 32(%ebp),%ebx
movl %eax,-8(%edi)
movl %edx,-4(%edi)
leal 32(%esi),%esi
leal 32(%ebp),%ebp
decl %ecx
jnz Loop
popl %edx
Lend:
decl %edx /* test %edx w/o clobbering carry */
js Lend2
incl %edx
Loop2:
leal 4(%edi),%edi
movl (%esi),%eax
sbbl %ebx,%eax
movl 4(%ebp),%ebx
movl %eax,-4(%edi)
leal 4(%esi),%esi
leal 4(%ebp),%ebp
decl %edx
jnz Loop2
Lend2:
movl (%esi),%eax
sbbl %ebx,%eax
movl %eax,(%edi)
sbbl %eax,%eax
negl %eax
popl %ebp
popl %ebx
popl %esi
popl %edi
ret

1499
mpi/longlong.h Normal file
View File

File diff suppressed because it is too large Load Diff

8
mpi/m68k/distfiles Normal file
View File

@ -0,0 +1,8 @@
syntax.h
mpih-lshift.S
mpih-rshift.S
mpih-add1.S
mpih-sub1.S

3
mpi/m68k/mc68020/distfiles Normal file
View File

@ -0,0 +1,3 @@
mpih-mul1.S
mpih-mul2.S
mpih-mul3.S

106
mpi/m68k/mc68020/mpih-mul1.S Normal file
View File

@ -0,0 +1,106 @@
/* mc68020 __mpn_mul_1 -- Multiply a limb vector with a limb and store
* the result in a second limb vector.
*
* Copyright (C) 1992, 1994, 1996, 1998,
* 2001 Free Software Foundation, Inc.
*
* This file is part of GnuPG.
*
* GnuPG is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* GnuPG is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA
*
* Note: This code is heavily based on the GNU MP Library.
* Actually it's the same code with only minor changes in the
* way the data is stored; this is to support the abstraction
* of an optional secure memory allocation which may be used
* to avoid revealing of sensitive data due to paging etc.
* The GNU MP Library itself is published under the LGPL;
* however I decided to publish this code under the plain GPL.
*/
#include "sysdep.h"
#include "asm-syntax.h"
/*******************
* mpi_limb_t
* mpihelp_mul_1( mpi_ptr_t res_ptr, (sp + 4)
* mpi_ptr_t s1_ptr, (sp + 8)
* mpi_size_t s1_size, (sp + 12)
* mpi_limb_t s2_limb) (sp + 16)
*/
TEXT
ALIGN
GLOBL C_SYMBOL_NAME(mpihelp_mul_1)
C_SYMBOL_NAME(mpihelp_mul_1:)
PROLOG(mpihelp_mul_1)
#define res_ptr a0
#define s1_ptr a1
#define s1_size d2
#define s2_limb d4
/* Save used registers on the stack. */
moveml R(d2)-R(d4),MEM_PREDEC(sp)
#if 0
movel R(d2),MEM_PREDEC(sp)
movel R(d3),MEM_PREDEC(sp)
movel R(d4),MEM_PREDEC(sp)
#endif
/* Copy the arguments to registers. Better use movem? */
movel MEM_DISP(sp,16),R(res_ptr)
movel MEM_DISP(sp,20),R(s1_ptr)
movel MEM_DISP(sp,24),R(s1_size)
movel MEM_DISP(sp,28),R(s2_limb)
eorw #1,R(s1_size)
clrl R(d1)
lsrl #1,R(s1_size)
bcc L(L1)
subql #1,R(s1_size)
subl R(d0),R(d0) /* (d0,cy) <= (0,0) */
L(Loop:)
movel MEM_POSTINC(s1_ptr),R(d3)
mulul R(s2_limb),R(d1):R(d3)
addxl R(d0),R(d3)
movel R(d3),MEM_POSTINC(res_ptr)
L(L1:) movel MEM_POSTINC(s1_ptr),R(d3)
mulul R(s2_limb),R(d0):R(d3)
addxl R(d1),R(d3)
movel R(d3),MEM_POSTINC(res_ptr)
dbf R(s1_size),L(Loop)
clrl R(d3)
addxl R(d3),R(d0)
subl #0x10000,R(s1_size)
bcc L(Loop)
/* Restore used registers from stack frame. */
moveml MEM_POSTINC(sp),R(d2)-R(d4)
#if 0
movel MEM_POSTINC(sp),R(d4)
movel MEM_POSTINC(sp),R(d3)
movel MEM_POSTINC(sp),R(d2)
#endif
rts
EPILOG(mpihelp_mul_1)

96
mpi/m68k/mc68020/mpih-mul2.S Normal file
View File

@ -0,0 +1,96 @@
/* mc68020 __mpn_addmul_1 -- Multiply a limb vector with a limb and add
* the result to a second limb vector.
*
* Copyright (C) 1992, 1994, 1996, 1998,
* 2001 Free Software Foundation, Inc.
*
* This file is part of GnuPG.
*
* GnuPG is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* GnuPG is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA
*
* Note: This code is heavily based on the GNU MP Library.
* Actually it's the same code with only minor changes in the
* way the data is stored; this is to support the abstraction
* of an optional secure memory allocation which may be used
* to avoid revealing of sensitive data due to paging etc.
* The GNU MP Library itself is published under the LGPL;
* however I decided to publish this code under the plain GPL.
*/
#include "sysdep.h"
#include "asm-syntax.h"
/*******************
* mpi_limb_t
* mpihelp_addmul_1( mpi_ptr_t res_ptr, (sp + 4)
* mpi_ptr_t s1_ptr, (sp + 8)
* mpi_size_t s1_size, (sp + 12)
* mpi_limb_t s2_limb) (sp + 16)
*/
TEXT
ALIGN
GLOBL C_SYMBOL_NAME(mpihelp_addmul_1)
C_SYMBOL_NAME(mpihelp_addmul_1:)
PROLOG(mpihelp_addmul_1)
#define res_ptr a0
#define s1_ptr a1
#define s1_size d2
#define s2_limb d4
/* Save used registers on the stack. */
moveml R(d2)-R(d5),MEM_PREDEC(sp)
/* Copy the arguments to registers. Better use movem? */
movel MEM_DISP(sp,20),R(res_ptr)
movel MEM_DISP(sp,24),R(s1_ptr)
movel MEM_DISP(sp,28),R(s1_size)
movel MEM_DISP(sp,32),R(s2_limb)
eorw #1,R(s1_size)
clrl R(d1)
clrl R(d5)
lsrl #1,R(s1_size)
bcc L(L1)
subql #1,R(s1_size)
subl R(d0),R(d0) /* (d0,cy) <= (0,0) */
L(Loop:)
movel MEM_POSTINC(s1_ptr),R(d3)
mulul R(s2_limb),R(d1):R(d3)
addxl R(d0),R(d3)
addxl R(d5),R(d1)
addl R(d3),MEM_POSTINC(res_ptr)
L(L1:) movel MEM_POSTINC(s1_ptr),R(d3)
mulul R(s2_limb),R(d0):R(d3)
addxl R(d1),R(d3)
addxl R(d5),R(d0)
addl R(d3),MEM_POSTINC(res_ptr)
dbf R(s1_size),L(Loop)
addxl R(d5),R(d0)
subl #0x10000,R(s1_size)
bcc L(Loop)
/* Restore used registers from stack frame. */
moveml MEM_POSTINC(sp),R(d2)-R(d5)
rts
EPILOG(mpihelp_addmul_1)

99
mpi/m68k/mc68020/mpih-mul3.S Normal file
View File

@ -0,0 +1,99 @@
/* mc68020 __mpn_submul_1 -- Multiply a limb vector with a limb and subtract
* the result from a second limb vector.
*
* Copyright (C) 1992, 1994, 1996, 1998,
* 2001 Free Software Foundation, Inc.
*
* This file is part of GnuPG.
*
* GnuPG is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* GnuPG is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA
*
* Note: This code is heavily based on the GNU MP Library.
* Actually it's the same code with only minor changes in the
* way the data is stored; this is to support the abstraction
* of an optional secure memory allocation which may be used
* to avoid revealing of sensitive data due to paging etc.
* The GNU MP Library itself is published under the LGPL;
* however I decided to publish this code under the plain GPL.
*/
#include "sysdep.h"
#include "asm-syntax.h"
/*******************
* mpi_limb_t
* mpihelp_submul_1( mpi_ptr_t res_ptr, (sp + 4)
* mpi_ptr_t s1_ptr, (sp + 8)
* mpi_size_t s1_size, (sp + 12)
* mpi_limb_t s2_limb) (sp + 16)
*/
TEXT
ALIGN
GLOBL C_SYMBOL_NAME(mpihelp_submul_1)
C_SYMBOL_NAME(mpihelp_submul_1:)
PROLOG(mpihelp_submul_1)
#define res_ptr a0
#define s1_ptr a1
#define s1_size d2
#define s2_limb d4
/* Save used registers on the stack. */
moveml R(d2)-R(d5),MEM_PREDEC(sp)
/* Copy the arguments to registers. Better use movem? */
movel MEM_DISP(sp,20),R(res_ptr)
movel MEM_DISP(sp,24),R(s1_ptr)
movel MEM_DISP(sp,28),R(s1_size)
movel MEM_DISP(sp,32),R(s2_limb)
eorw #1,R(s1_size)
clrl R(d1)
clrl R(d5)
lsrl #1,R(s1_size)
bcc L(L1)
subql #1,R(s1_size)
subl R(d0),R(d0) /* (d0,cy) <= (0,0) */
L(Loop:)
movel MEM_POSTINC(s1_ptr),R(d3)
mulul R(s2_limb),R(d1):R(d3)
addxl R(d0),R(d3)
addxl R(d5),R(d1)
subl R(d3),MEM_POSTINC(res_ptr)
L(L1:) movel MEM_POSTINC(s1_ptr),R(d3)
mulul R(s2_limb),R(d0):R(d3)
addxl R(d1),R(d3)
addxl R(d5),R(d0)
subl R(d3),MEM_POSTINC(res_ptr)
dbf R(s1_size),L(Loop)
addxl R(d5),R(d0)
subl #0x10000,R(s1_size)
bcc L(Loop)
/* Restore used registers from stack frame. */
moveml MEM_POSTINC(sp),R(d2)-R(d5)
rts
EPILOG(mpihelp_submul_1)

94
mpi/m68k/mpih-add1.S Normal file
View File

@ -0,0 +1,94 @@
/* mc68020 __mpn_add_n -- Add two limb vectors of the same length > 0 and store
* sum in a third limb vector.
*
* Copyright (C) 1992, 1994,1996, 1998,
* 2001 Free Software Foundation, Inc.
*
* This file is part of GnuPG.
*
* GnuPG is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* GnuPG is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA
*
* Note: This code is heavily based on the GNU MP Library.
* Actually it's the same code with only minor changes in the
* way the data is stored; this is to support the abstraction
* of an optional secure memory allocation which may be used
* to avoid revealing of sensitive data due to paging etc.
* The GNU MP Library itself is published under the LGPL;
* however I decided to publish this code under the plain GPL.
*/
#include "sysdep.h"
#include "asm-syntax.h"
/*******************
* mpi_limb_t
* mpihelp_add_n( mpi_ptr_t res_ptr, (sp + 4)
* mpi_ptr_t s1_ptr, (sp + 8)
* mpi_ptr_t s2_ptr, (sp + 16)
* mpi_size_t size) (sp + 12)
*/
TEXT
ALIGN
GLOBL C_SYMBOL_NAME(mpihelp_add_n)
C_SYMBOL_NAME(mpihelp_add_n:)
PROLOG(mpihelp_add_n)
/* Save used registers on the stack. */
movel R(d2),MEM_PREDEC(sp)
movel R(a2),MEM_PREDEC(sp)
/* Copy the arguments to registers. Better use movem? */
movel MEM_DISP(sp,12),R(a2)
movel MEM_DISP(sp,16),R(a0)
movel MEM_DISP(sp,20),R(a1)
movel MEM_DISP(sp,24),R(d2)
eorw #1,R(d2)
lsrl #1,R(d2)
bcc L(L1)
subql #1,R(d2) /* clears cy as side effect */
L(Loop:)
movel MEM_POSTINC(a0),R(d0)
movel MEM_POSTINC(a1),R(d1)
addxl R(d1),R(d0)
movel R(d0),MEM_POSTINC(a2)
L(L1:) movel MEM_POSTINC(a0),R(d0)
movel MEM_POSTINC(a1),R(d1)
addxl R(d1),R(d0)
movel R(d0),MEM_POSTINC(a2)
dbf R(d2),L(Loop) /* loop until 16 lsb of %4 == -1 */
subxl R(d0),R(d0) /* d0 <= -cy; save cy as 0 or -1 in d0 */
subl #0x10000,R(d2)
bcs L(L2)
addl R(d0),R(d0) /* restore cy */
bra L(Loop)
L(L2:)
negl R(d0)
/* Restore used registers from stack frame. */
movel MEM_POSTINC(sp),R(a2)
movel MEM_POSTINC(sp),R(d2)
rts
EPILOG(mpihelp_add_n)

166
mpi/m68k/mpih-lshift.S Normal file
View File

@ -0,0 +1,166 @@
/* mc68020 lshift -- Shift left a low-level natural-number integer.
*
* Copyright (C) 1996, 1998, 2001 Free Software Foundation, Inc.
*
* This file is part of GnuPG.
*
* GnuPG is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* GnuPG is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA
*
* Note: This code is heavily based on the GNU MP Library.
* Actually it's the same code with only minor changes in the
* way the data is stored; this is to support the abstraction
* of an optional secure memory allocation which may be used
* to avoid revealing of sensitive data due to paging etc.
* The GNU MP Library itself is published under the LGPL;
* however I decided to publish this code under the plain GPL.
*/
#include "sysdep.h"
#include "asm-syntax.h"
/*******************
* mpi_limb_t
* mpihelp_lshift( mpi_ptr_t wp, (sp + 4)
* mpi_ptr_t up, (sp + 8)
* mpi_size_t usize, (sp + 12)
* unsigned cnt) (sp + 16)
*/
#define res_ptr a1
#define s_ptr a0
#define s_size d6
#define cnt d4
TEXT
ALIGN
GLOBL C_SYMBOL_NAME(mpihelp_lshift)
C_SYMBOL_NAME(mpihelp_lshift:)
PROLOG(mpihelp_lshift)
/* Save used registers on the stack. */
moveml R(d2)-R(d6)/R(a2),MEM_PREDEC(sp)
/* Copy the arguments to registers. */
movel MEM_DISP(sp,28),R(res_ptr)
movel MEM_DISP(sp,32),R(s_ptr)
movel MEM_DISP(sp,36),R(s_size)
movel MEM_DISP(sp,40),R(cnt)
moveql #1,R(d5)
cmpl R(d5),R(cnt)
bne L(Lnormal)
cmpl R(s_ptr),R(res_ptr)
bls L(Lspecial) /* jump if s_ptr >= res_ptr */
#if (defined (__mc68020__) || defined (__NeXT__) || defined(mc68020))
lea MEM_INDX1(s_ptr,s_size,l,4),R(a2)
#else /* not mc68020 */
movel R(s_size),R(d0)
asll #2,R(d0)
lea MEM_INDX(s_ptr,d0,l),R(a2)
#endif
cmpl R(res_ptr),R(a2)
bls L(Lspecial) /* jump if res_ptr >= s_ptr + s_size */
L(Lnormal:)
moveql #32,R(d5)
subl R(cnt),R(d5)
#if (defined (__mc68020__) || defined (__NeXT__) || defined(mc68020))
lea MEM_INDX1(s_ptr,s_size,l,4),R(s_ptr)
lea MEM_INDX1(res_ptr,s_size,l,4),R(res_ptr)
#else /* not mc68000 */
movel R(s_size),R(d0)
asll #2,R(d0)
addl R(s_size),R(s_ptr)
addl R(s_size),R(res_ptr)
#endif
movel MEM_PREDEC(s_ptr),R(d2)
movel R(d2),R(d0)
lsrl R(d5),R(d0) /* compute carry limb */
lsll R(cnt),R(d2)
movel R(d2),R(d1)
subql #1,R(s_size)
beq L(Lend)
lsrl #1,R(s_size)
bcs L(L1)
subql #1,R(s_size)
L(Loop:)
movel MEM_PREDEC(s_ptr),R(d2)
movel R(d2),R(d3)
lsrl R(d5),R(d3)
orl R(d3),R(d1)
movel R(d1),MEM_PREDEC(res_ptr)
lsll R(cnt),R(d2)
L(L1:)
movel MEM_PREDEC(s_ptr),R(d1)
movel R(d1),R(d3)
lsrl R(d5),R(d3)
orl R(d3),R(d2)
movel R(d2),MEM_PREDEC(res_ptr)
lsll R(cnt),R(d1)
dbf R(s_size),L(Loop)
subl #0x10000,R(s_size)
bcc L(Loop)
L(Lend:)
movel R(d1),MEM_PREDEC(res_ptr) /* store least significant limb */
/* Restore used registers from stack frame. */
moveml MEM_POSTINC(sp),R(d2)-R(d6)/R(a2)
rts
/* We loop from least significant end of the arrays, which is only
permissable if the source and destination don't overlap, since the
function is documented to work for overlapping source and destination. */
L(Lspecial:)
clrl R(d0) /* initialize carry */
eorw #1,R(s_size)
lsrl #1,R(s_size)
bcc L(LL1)
subql #1,R(s_size)
L(LLoop:)
movel MEM_POSTINC(s_ptr),R(d2)
addxl R(d2),R(d2)
movel R(d2),MEM_POSTINC(res_ptr)
L(LL1:)
movel MEM_POSTINC(s_ptr),R(d2)
addxl R(d2),R(d2)
movel R(d2),MEM_POSTINC(res_ptr)
dbf R(s_size),L(LLoop)
addxl R(d0),R(d0) /* save cy in lsb */
subl #0x10000,R(s_size)
bcs L(LLend)
lsrl #1,R(d0) /* restore cy */
bra L(LLoop)
L(LLend:)
/* Restore used registers from stack frame. */
moveml MEM_POSTINC(sp),R(d2)-R(d6)/R(a2)
rts
EPILOG(mpihelp_lshift)

164
mpi/m68k/mpih-rshift.S Normal file
View File

@ -0,0 +1,164 @@
/* mc68020 rshift -- Shift right a low-level natural-number integer.
*
* Copyright (C) 1996, 1998, 2001 Free Software Foundation, Inc.
*
* This file is part of GnuPG.
*
* GnuPG is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* GnuPG is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA
*
* Note: This code is heavily based on the GNU MP Library.
* Actually it's the same code with only minor changes in the
* way the data is stored; this is to support the abstraction
* of an optional secure memory allocation which may be used
* to avoid revealing of sensitive data due to paging etc.
* The GNU MP Library itself is published under the LGPL;
* however I decided to publish this code under the plain GPL.
*/
#include "sysdep.h"
#include "asm-syntax.h"
/*******************
* mpi_limb_t
* mpihelp_rshift( mpi_ptr_t wp, (sp + 4)
* mpi_ptr_t up, (sp + 8)
* mpi_size_t usize, (sp + 12)
* unsigned cnt) (sp + 16)
*/
#define res_ptr a1
#define s_ptr a0
#define s_size d6
#define cnt d4
TEXT
ALIGN
GLOBL C_SYMBOL_NAME(mpihelp_rshift)
C_SYMBOL_NAME(mpihelp_rshift:)
PROLOG(mpihelp_rshift)
/* Save used registers on the stack. */
moveml R(d2)-R(d6)/R(a2),MEM_PREDEC(sp)
/* Copy the arguments to registers. */
movel MEM_DISP(sp,28),R(res_ptr)
movel MEM_DISP(sp,32),R(s_ptr)
movel MEM_DISP(sp,36),R(s_size)
movel MEM_DISP(sp,40),R(cnt)
moveql #1,R(d5)
cmpl R(d5),R(cnt)
bne L(Rnormal)
cmpl R(res_ptr),R(s_ptr)
bls L(Rspecial) /* jump if res_ptr >= s_ptr */
#if (defined (__mc68020__) || defined (__NeXT__) || defined(mc68020))
lea MEM_INDX1(res_ptr,s_size,l,4),R(a2)
#else /* not mc68020 */
movel R(s_size),R(d0)
asll #2,R(d0)
lea MEM_INDX(res_ptr,d0,l),R(a2)
#endif
cmpl R(s_ptr),R(a2)
bls L(Rspecial) /* jump if s_ptr >= res_ptr + s_size */
L(Rnormal:)
moveql #32,R(d5)
subl R(cnt),R(d5)
movel MEM_POSTINC(s_ptr),R(d2)
movel R(d2),R(d0)
lsll R(d5),R(d0) /* compute carry limb */
lsrl R(cnt),R(d2)
movel R(d2),R(d1)
subql #1,R(s_size)
beq L(Rend)
lsrl #1,R(s_size)
bcs L(R1)
subql #1,R(s_size)
L(Roop:)
movel MEM_POSTINC(s_ptr),R(d2)
movel R(d2),R(d3)
lsll R(d5),R(d3)
orl R(d3),R(d1)
movel R(d1),MEM_POSTINC(res_ptr)
lsrl R(cnt),R(d2)
L(R1:)
movel MEM_POSTINC(s_ptr),R(d1)
movel R(d1),R(d3)
lsll R(d5),R(d3)
orl R(d3),R(d2)
movel R(d2),MEM_POSTINC(res_ptr)
lsrl R(cnt),R(d1)
dbf R(s_size),L(Roop)
subl #0x10000,R(s_size)
bcc L(Roop)
L(Rend:)
movel R(d1),MEM(res_ptr) /* store most significant limb */
/* Restore used registers from stack frame. */
moveml MEM_POSTINC(sp),R(d2)-R(d6)/R(a2)
rts
/* We loop from most significant end of the arrays, which is only
permissable if the source and destination don't overlap, since the
function is documented to work for overlapping source and destination. */
L(Rspecial:)
#if (defined (__mc68020__) || defined (__NeXT__) || defined(mc68020))
lea MEM_INDX1(s_ptr,s_size,l,4),R(s_ptr)
lea MEM_INDX1(res_ptr,s_size,l,4),R(res_ptr)
#else /* not mc68000 */
movel R(s_size),R(d0)
asll #2,R(d0)
addl R(s_size),R(s_ptr)
addl R(s_size),R(res_ptr)
#endif
clrl R(d0) /* initialize carry */
eorw #1,R(s_size)
lsrl #1,R(s_size)
bcc L(LR1)
subql #1,R(s_size)
L(LRoop:)
movel MEM_PREDEC(s_ptr),R(d2)
roxrl #1,R(d2)
movel R(d2),MEM_PREDEC(res_ptr)
L(LR1:)
movel MEM_PREDEC(s_ptr),R(d2)
roxrl #1,R(d2)
movel R(d2),MEM_PREDEC(res_ptr)
dbf R(s_size),L(LRoop)
roxrl #1,R(d0) /* save cy in msb */
subl #0x10000,R(s_size)
bcs L(LRend)
addl R(d0),R(d0) /* restore cy */
bra L(LRoop)
L(LRend:)
/* Restore used registers from stack frame. */
moveml MEM_POSTINC(sp),R(d2)-R(d6)/R(a2)
rts
EPILOG(mpihelp_rshift)

93
mpi/m68k/mpih-sub1.S Normal file
View File

@ -0,0 +1,93 @@
/* mc68020 __mpn_sub_n -- Subtract two limb vectors of the same length > 0 and
* store difference in a third limb vector.
*
* Copyright (C) 1992, 1994, 1996, 1998,
* 2001 Free Software Foundation, Inc.
*
* This file is part of GnuPG.
*
* GnuPG is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* GnuPG is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA
*
* Note: This code is heavily based on the GNU MP Library.
* Actually it's the same code with only minor changes in the
* way the data is stored; this is to support the abstraction
* of an optional secure memory allocation which may be used
* to avoid revealing of sensitive data due to paging etc.
* The GNU MP Library itself is published under the LGPL;
* however I decided to publish this code under the plain GPL.
*/
#include "sysdep.h"
#include "asm-syntax.h"
/*******************
* mpi_limb_t
* mpihelp_sub_n( mpi_ptr_t res_ptr, (sp + 4)
* mpi_ptr_t s1_ptr, (sp + 8)
* mpi_ptr_t s2_ptr, (sp + 16)
* mpi_size_t size) (sp + 12)
*/
TEXT
ALIGN
GLOBL C_SYMBOL_NAME(mpihelp_sub_n)
C_SYMBOL_NAME(mpihelp_sub_n:)
PROLOG(mpihelp_sub_n)
/* Save used registers on the stack. */
movel R(d2),MEM_PREDEC(sp)
movel R(a2),MEM_PREDEC(sp)
/* Copy the arguments to registers. Better use movem? */
movel MEM_DISP(sp,12),R(a2)
movel MEM_DISP(sp,16),R(a0)
movel MEM_DISP(sp,20),R(a1)
movel MEM_DISP(sp,24),R(d2)
eorw #1,R(d2)
lsrl #1,R(d2)
bcc L(L1)
subql #1,R(d2) /* clears cy as side effect */
L(Loop:)
movel MEM_POSTINC(a0),R(d0)
movel MEM_POSTINC(a1),R(d1)
subxl R(d1),R(d0)
movel R(d0),MEM_POSTINC(a2)
L(L1:) movel MEM_POSTINC(a0),R(d0)
movel MEM_POSTINC(a1),R(d1)
subxl R(d1),R(d0)
movel R(d0),MEM_POSTINC(a2)
dbf R(d2),L(Loop) /* loop until 16 lsb of %4 == -1 */
subxl R(d0),R(d0) /* d0 <= -cy; save cy as 0 or -1 in d0 */
subl #0x10000,R(d2)
bcs L(L2)
addl R(d0),R(d0) /* restore cy */
bra L(Loop)
L(L2:)
negl R(d0)
/* Restore used registers from stack frame. */
movel MEM_POSTINC(sp),R(a2)
movel MEM_POSTINC(sp),R(d2)
rts
EPILOG(mpihelp_sub_n)

187
mpi/m68k/syntax.h Normal file
View File

@ -0,0 +1,187 @@
/* asm.h -- Definitions for 68k syntax variations.
*
* Copyright (C) 1992, 1994, 1996, 1998,
* 2001 Free Software Foundation, Inc.
*
* This file is part of GnuPG.
*
* GnuPG is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* GnuPG is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA
*
* Note: This code is heavily based on the GNU MP Library.
* Actually it's the same code with only minor changes in the
* way the data is stored; this is to support the abstraction
* of an optional secure memory allocation which may be used
* to avoid revealing of sensitive data due to paging etc.
* The GNU MP Library itself is published under the LGPL;
* however I decided to publish this code under the plain GPL.
*/
#undef ALIGN
#ifdef MIT_SYNTAX
#define PROLOG(name)
#define EPILOG(name)
#define R(r)r
#define MEM(base)base@
#define MEM_DISP(base,displacement)base@(displacement)
#define MEM_INDX(base,idx,size_suffix)base@(idx:size_suffix)
#define MEM_INDX1(base,idx,size_suffix,scale)base@(idx:size_suffix:scale)
#define MEM_PREDEC(memory_base)memory_base@-
#define MEM_POSTINC(memory_base)memory_base@+
#define L(label) label
#define TEXT .text
#define ALIGN .even
#define GLOBL .globl
#define moveql moveq
/* Use variable sized opcodes. */
#define bcc jcc
#define bcs jcs
#define bls jls
#define beq jeq
#define bne jne
#define bra jra
#endif
#ifdef SONY_SYNTAX
#define PROLOG(name)
#define EPILOG(name)
#define R(r)r
#define MEM(base)(base)
#define MEM_DISP(base,displacement)(displacement,base)
#define MEM_INDX(base,idx,size_suffix)(base,idx.size_suffix)
#define MEM_INDX1(base,idx,size_suffix,scale)(base,idx.size_suffix*scale)
#define MEM_PREDEC(memory_base)-(memory_base)
#define MEM_POSTINC(memory_base)(memory_base)+
#define L(label) label
#define TEXT .text
#define ALIGN .even
#define GLOBL .globl
#endif
#ifdef MOTOROLA_SYNTAX
#define PROLOG(name)
#define EPILOG(name)
#define R(r)r
#define MEM(base)(base)
#define MEM_DISP(base,displacement)(displacement,base)
#define MEM_INDX(base,idx,size_suffix)(base,idx.size_suffix)
#define MEM_INDX1(base,idx,size_suffix,scale)(base,idx.size_suffix*scale)
#define MEM_PREDEC(memory_base)-(memory_base)
#define MEM_POSTINC(memory_base)(memory_base)+
#define L(label) label
#define TEXT
#define ALIGN
#define GLOBL XDEF
#define lea LEA
#define movel MOVE.L
#define moveml MOVEM.L
#define moveql MOVEQ.L
#define cmpl CMP.L
#define orl OR.L
#define clrl CLR.L
#define eorw EOR.W
#define lsrl LSR.L
#define lsll LSL.L
#define roxrl ROXR.L
#define roxll ROXL.L
#define addl ADD.L
#define addxl ADDX.L
#define addql ADDQ.L
#define subl SUB.L
#define subxl SUBX.L
#define subql SUBQ.L
#define negl NEG.L
#define mulul MULU.L
#define bcc BCC
#define bcs BCS
#define bls BLS
#define beq BEQ
#define bne BNE
#define bra BRA
#define dbf DBF
#define rts RTS
#define d0 D0
#define d1 D1
#define d2 D2
#define d3 D3
#define d4 D4
#define d5 D5
#define d6 D6
#define d7 D7
#define a0 A0
#define a1 A1
#define a2 A2
#define a3 A3
#define a4 A4
#define a5 A5
#define a6 A6
#define a7 A7
#define sp SP
#endif
#ifdef ELF_SYNTAX
#define PROLOG(name) .type name,@function
#define EPILOG(name) .size name,.-name
#define MEM(base)(R(base))
#define MEM_DISP(base,displacement)(displacement,R(base))
#define MEM_PREDEC(memory_base)-(R(memory_base))
#define MEM_POSTINC(memory_base)(R(memory_base))+
#ifdef __STDC__
#define R_(r)%##r
#define R(r)R_(r)
#define MEM_INDX_(base,idx,size_suffix)(R(base),R(idx##.##size_suffix))
#define MEM_INDX(base,idx,size_suffix)MEM_INDX_(base,idx,size_suffix)
#define MEM_INDX1_(base,idx,size_suffix,scale)(R(base),R(idx##.##size_suffix*scale))
#define MEM_INDX1(base,idx,size_suffix,scale)MEM_INDX1_(base,idx,size_suffix,scale)
#define L(label) .##label
#else
#define R(r)%/**/r
#define MEM_INDX(base,idx,size_suffix)(R(base),R(idx).size_suffix)
#define MEM_INDX1(base,idx,size_suffix,scale)(R(base),R(idx).size_suffix*scale)
#define L(label) ./**/label
#endif
#define TEXT .text
#define ALIGN .align 2
#define GLOBL .globl
#define bcc jbcc
#define bcs jbcs
#define bls jbls
#define beq jbeq
#define bne jbne
#define bra jbra
#endif
#if defined (SONY_SYNTAX) || defined (ELF_SYNTAX)
#define movel move.l
#define moveml movem.l
#define moveql moveq.l
#define cmpl cmp.l
#define orl or.l
#define clrl clr.l
#define eorw eor.w
#define lsrl lsr.l
#define lsll lsl.l
#define roxrl roxr.l
#define roxll roxl.l
#define addl add.l
#define addxl addx.l
#define addql addq.l
#define subl sub.l
#define subxl subx.l
#define subql subq.l
#define negl neg.l
#define mulul mulu.l
#endif

23
mpi/mips3/README Normal file
View File

@ -0,0 +1,23 @@
This directory contains mpn functions optimized for MIPS3. Example of
processors that implement MIPS3 are R4000, R4400, R4600, R4700, and R8000.
RELEVANT OPTIMIZATION ISSUES
1. On the R4000 and R4400, branches, both the plain and the "likely" ones,
take 3 cycles to execute. (The fastest possible loop will take 4 cycles,
because of the delay insn.)
On the R4600, branches takes a single cycle
On the R8000, branches often take no noticable cycles, as they are
executed in a separate function unit..
2. The R4000 and R4400 have a load latency of 4 cycles.
3. On the R4000 and R4400, multiplies take a data-dependent number of
cycles, contrary to the SGI documentation. There seem to be 3 or 4
possible latencies.
STATUS
Good...

10
mpi/mips3/distfiles Normal file
View File

@ -0,0 +1,10 @@
README
mpih-add1.S
mpih-sub1.S
mpih-mul1.S
mpih-mul2.S
mpih-mul3.S
mpih-lshift.S
mpih-rshift.S
mpi-asm-defs.h

10
mpi/mips3/mpi-asm-defs.h Normal file
View File

@ -0,0 +1,10 @@
/* This file defines some basic constants for the MPI machinery. We
* need to define the types on a per-CPU basis, so it is done with
* this file here. */
#define BYTES_PER_MPI_LIMB 8

124
mpi/mips3/mpih-add1.S Normal file
View File

@ -0,0 +1,124 @@
/* mips3 add_n -- Add two limb vectors of the same length > 0 and store
* sum in a third limb vector.
*
* Copyright (C) 1995, 1998, 2000
* 2001 Free Software Foundation, Inc.
*
* This file is part of GnuPG.
*
* GnuPG is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* GnuPG is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA
*/
/*******************
* mpi_limb_t
* mpihelp_add_n( mpi_ptr_t res_ptr, ($4)
* mpi_ptr_t s1_ptr, ($5)
* mpi_ptr_t s2_ptr, ($6)
* mpi_size_t size) ($7)
*/
.text
.align 2
.globl mpihelp_add_n
.ent mpihelp_add_n
mpihelp_add_n:
.set noreorder
.set nomacro
ld $10,0($5)
ld $11,0($6)
daddiu $7,$7,-1
and $9,$7,4-1 # number of limbs in first loop
beq $9,$0,.L0 # if multiple of 4 limbs, skip first loop
move $2,$0
dsubu $7,$7,$9
.Loop0: daddiu $9,$9,-1
ld $12,8($5)
daddu $11,$11,$2
ld $13,8($6)
sltu $8,$11,$2
daddu $11,$10,$11
sltu $2,$11,$10
sd $11,0($4)
or $2,$2,$8
daddiu $5,$5,8
daddiu $6,$6,8
move $10,$12
move $11,$13
bne $9,$0,.Loop0
daddiu $4,$4,8
.L0: beq $7,$0,.Lend
nop
.Loop: daddiu $7,$7,-4
ld $12,8($5)
daddu $11,$11,$2
ld $13,8($6)
sltu $8,$11,$2
daddu $11,$10,$11
sltu $2,$11,$10
sd $11,0($4)
or $2,$2,$8
ld $10,16($5)
daddu $13,$13,$2
ld $11,16($6)
sltu $8,$13,$2
daddu $13,$12,$13
sltu $2,$13,$12
sd $13,8($4)
or $2,$2,$8
ld $12,24($5)
daddu $11,$11,$2
ld $13,24($6)
sltu $8,$11,$2
daddu $11,$10,$11
sltu $2,$11,$10
sd $11,16($4)
or $2,$2,$8
ld $10,32($5)
daddu $13,$13,$2
ld $11,32($6)
sltu $8,$13,$2
daddu $13,$12,$13
sltu $2,$13,$12
sd $13,24($4)
or $2,$2,$8
daddiu $5,$5,32
daddiu $6,$6,32
bne $7,$0,.Loop
daddiu $4,$4,32
.Lend: daddu $11,$11,$2
sltu $8,$11,$2
daddu $11,$10,$11
sltu $2,$11,$10
sd $11,0($4)
j $31
or $2,$2,$8
.end mpihelp_add_n

97
mpi/mips3/mpih-lshift.S Normal file
View File

@ -0,0 +1,97 @@
/* mips3 lshift
*
* Copyright (C) 1995, 1998, 2000,
* 2001 Free Software Foundation, Inc.
*
* This file is part of GnuPG.
*
* GnuPG is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* GnuPG is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA
*/
/*******************
* mpi_limb_t
* mpihelp_lshift( mpi_ptr_t wp, ($4)
* mpi_ptr_t up, ($5)
* mpi_size_t usize, ($6)
* unsigned cnt) ($7)
*/
.text
.align 2
.globl mpihelp_lshift
.ent mpihelp_lshift
mpihelp_lshift:
.set noreorder
.set nomacro
dsll $2,$6,3
daddu $5,$5,$2 # make r5 point at end of src
ld $10,-8($5) # load first limb
dsubu $13,$0,$7
daddu $4,$4,$2 # make r4 point at end of res
daddiu $6,$6,-1
and $9,$6,4-1 # number of limbs in first loop
beq $9,$0,.L0 # if multiple of 4 limbs, skip first loop
dsrl $2,$10,$13 # compute function result
dsubu $6,$6,$9
.Loop0: ld $3,-16($5)
daddiu $4,$4,-8
daddiu $5,$5,-8
daddiu $9,$9,-1
dsll $11,$10,$7
dsrl $12,$3,$13
move $10,$3
or $8,$11,$12
bne $9,$0,.Loop0
sd $8,0($4)
.L0: beq $6,$0,.Lend
nop
.Loop: ld $3,-16($5)
daddiu $4,$4,-32
daddiu $6,$6,-4
dsll $11,$10,$7
dsrl $12,$3,$13
ld $10,-24($5)
dsll $14,$3,$7
or $8,$11,$12
sd $8,24($4)
dsrl $9,$10,$13
ld $3,-32($5)
dsll $11,$10,$7
or $8,$14,$9
sd $8,16($4)
dsrl $12,$3,$13
ld $10,-40($5)
dsll $14,$3,$7
or $8,$11,$12
sd $8,8($4)
dsrl $9,$10,$13
daddiu $5,$5,-32
or $8,$14,$9
bgtz $6,.Loop
sd $8,0($4)
.Lend: dsll $8,$10,$7
j $31
sd $8,-8($4)
.end mpihelp_lshift

89
mpi/mips3/mpih-mul1.S Normal file
View File

@ -0,0 +1,89 @@
/* mips3 mpih-mul1.S -- Multiply a limb vector with a limb and store
* the result in a second limb vector.
*
* Copyright (C) 1992, 1994, 1995, 1998, 2000
* 2001 Free Software Foundation, Inc.
*
* This file is part of GnuPG.
*
* GnuPG is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* GnuPG is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA
*/
/*******************
* mpi_limb_t
* mpihelp_mul_1( mpi_ptr_t res_ptr, (r4)
* mpi_ptr_t s1_ptr, (r5)
* mpi_size_t s1_size, (r6)
* mpi_limb_t s2_limb) (r7)
*/
.text
.align 4
.globl mpihelp_mul_1
.ent mpihelp_mul_1
mpihelp_mul_1:
.set noreorder
.set nomacro
/* # warm up phase 0 */
ld $8,0($5)
/* # warm up phase 1 */
daddiu $5,$5,8
dmultu $8,$7
daddiu $6,$6,-1
beq $6,$0,$LC0
move $2,$0 # zero cy2
daddiu $6,$6,-1
beq $6,$0,$LC1
ld $8,0($5) # load new s1 limb as early as possible
Loop: mflo $10
mfhi $9
daddiu $5,$5,8
daddu $10,$10,$2 # add old carry limb to low product limb
dmultu $8,$7
ld $8,0($5) # load new s1 limb as early as possible
daddiu $6,$6,-1 # decrement loop counter
sltu $2,$10,$2 # carry from previous addition -> $2
sd $10,0($4)
daddiu $4,$4,8
bne $6,$0,Loop
daddu $2,$9,$2 # add high product limb and carry from addition
/* # cool down phase 1 */
$LC1: mflo $10
mfhi $9
daddu $10,$10,$2
sltu $2,$10,$2
dmultu $8,$7
sd $10,0($4)
daddiu $4,$4,8
daddu $2,$9,$2 # add high product limb and carry from addition
/* # cool down phase 0 */
$LC0: mflo $10
mfhi $9
daddu $10,$10,$2
sltu $2,$10,$2
sd $10,0($4)
j $31
daddu $2,$9,$2 # add high product limb and carry from addition
.end mpihelp_mul_1

101
mpi/mips3/mpih-mul2.S Normal file
View File

@ -0,0 +1,101 @@
/* MIPS3 addmul_1 -- Multiply a limb vector with a single limb and
* add the product to a second limb vector.
*
* Copyright (C) 1992, 1994, 1995, 1998, 2000
* 2001 Free Software Foundation, Inc.
*
* This file is part of GnuPG.
*
* GnuPG is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* GnuPG is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA
*/
/*******************
* mpi_limb_t
* mpihelp_addmul_1( mpi_ptr_t res_ptr, (r4)
* mpi_ptr_t s1_ptr, (r5)
* mpi_size_t s1_size, (r6)
* mpi_limb_t s2_limb) (r7)
*/
.text
.align 4
.globl mpihelp_addmul_1
.ent mpihelp_addmul_1
mpihelp_addmul_1:
.set noreorder
.set nomacro
/* # warm up phase 0 */
ld $8,0($5)
/* # warm up phase 1 */
daddiu $5,$5,8
dmultu $8,$7
daddiu $6,$6,-1
beq $6,$0,$LC0
move $2,$0 # zero cy2
daddiu $6,$6,-1
beq $6,$0,$LC1
ld $8,0($5) # load new s1 limb as early as possible
Loop: ld $10,0($4)
mflo $3
mfhi $9
daddiu $5,$5,8
daddu $3,$3,$2 # add old carry limb to low product limb
dmultu $8,$7
ld $8,0($5) # load new s1 limb as early as possible
daddiu $6,$6,-1 # decrement loop counter
sltu $2,$3,$2 # carry from previous addition -> $2
daddu $3,$10,$3
sltu $10,$3,$10
daddu $2,$2,$10
sd $3,0($4)
daddiu $4,$4,8
bne $6,$0,Loop
daddu $2,$9,$2 # add high product limb and carry from addition
/* # cool down phase 1 */
$LC1: ld $10,0($4)
mflo $3
mfhi $9
daddu $3,$3,$2
sltu $2,$3,$2
dmultu $8,$7
daddu $3,$10,$3
sltu $10,$3,$10
daddu $2,$2,$10
sd $3,0($4)
daddiu $4,$4,8
daddu $2,$9,$2 # add high product limb and carry from addition
/* # cool down phase 0 */
$LC0: ld $10,0($4)
mflo $3
mfhi $9
daddu $3,$3,$2
sltu $2,$3,$2
daddu $3,$10,$3
sltu $10,$3,$10
daddu $2,$2,$10
sd $3,0($4)
j $31
daddu $2,$9,$2 # add high product limb and carry from addition
.end mpihelp_addmul_1

101
mpi/mips3/mpih-mul3.S Normal file
View File

@ -0,0 +1,101 @@
/* MIPS3 submul_1 -- Multiply a limb vector with a single limb and
* subtract the product from a second limb vector.
*
* Copyright (C) 1992, 1994, 1995, 1998, 2000
* 2001 Free Software Foundation, Inc.
*
* This file is part of GnuPG.
*
* GnuPG is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* GnuPG is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA
*/
/*******************
* mpi_limb_t
* mpihelp_submul_1( mpi_ptr_t res_ptr, (r4)
* mpi_ptr_t s1_ptr, (r5)
* mpi_size_t s1_size, (r6)
* mpi_limb_t s2_limb) (r7)
*/
.text
.align 4
.globl mpihelp_submul_1
.ent mpihelp_submul_1
mpihelp_submul_1:
.set noreorder
.set nomacro
/* # warm up phase 0 */
ld $8,0($5)
/* # warm up phase 1 */
daddiu $5,$5,8
dmultu $8,$7
daddiu $6,$6,-1
beq $6,$0,$LC0
move $2,$0 # zero cy2
daddiu $6,$6,-1
beq $6,$0,$LC1
ld $8,0($5) # load new s1 limb as early as possible
Loop: ld $10,0($4)
mflo $3
mfhi $9
daddiu $5,$5,8
daddu $3,$3,$2 # add old carry limb to low product limb
dmultu $8,$7
ld $8,0($5) # load new s1 limb as early as possible
daddiu $6,$6,-1 # decrement loop counter
sltu $2,$3,$2 # carry from previous addition -> $2
dsubu $3,$10,$3
sgtu $10,$3,$10
daddu $2,$2,$10
sd $3,0($4)
daddiu $4,$4,8
bne $6,$0,Loop
daddu $2,$9,$2 # add high product limb and carry from addition
/* # cool down phase 1 */
$LC1: ld $10,0($4)
mflo $3
mfhi $9
daddu $3,$3,$2
sltu $2,$3,$2
dmultu $8,$7
dsubu $3,$10,$3
sgtu $10,$3,$10
daddu $2,$2,$10
sd $3,0($4)
daddiu $4,$4,8
daddu $2,$9,$2 # add high product limb and carry from addition
/* # cool down phase 0 */
$LC0: ld $10,0($4)
mflo $3
mfhi $9
daddu $3,$3,$2
sltu $2,$3,$2
dsubu $3,$10,$3
sgtu $10,$3,$10
daddu $2,$2,$10
sd $3,0($4)
j $31
daddu $2,$9,$2 # add high product limb and carry from addition
.end mpihelp_submul_1

95
mpi/mips3/mpih-rshift.S Normal file
View File

@ -0,0 +1,95 @@
/* mips3 rshift
*
* Copyright (C) 1995, 1998, 2000
* 2001 Free Software Foundation, Inc.
*
* This file is part of GnuPG.
*
* GnuPG is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* GnuPG is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA
*/
/*******************
* mpi_limb_t
* mpihelp_rshift( mpi_ptr_t wp, ($4)
* mpi_ptr_t up, ($5)
* mpi_size_t usize, ($6)
* unsigned cnt) ($7)
*/
.text
.align 2
.globl mpihelp_rshift
.ent mpihelp_rshift
mpihelp_rshift:
.set noreorder
.set nomacro
ld $10,0($5) # load first limb
dsubu $13,$0,$7
daddiu $6,$6,-1
and $9,$6,4-1 # number of limbs in first loop
beq $9,$0,.L0 # if multiple of 4 limbs, skip first loop
dsll $2,$10,$13 # compute function result
dsubu $6,$6,$9
.Loop0: ld $3,8($5)
daddiu $4,$4,8
daddiu $5,$5,8
daddiu $9,$9,-1
dsrl $11,$10,$7
dsll $12,$3,$13
move $10,$3
or $8,$11,$12
bne $9,$0,.Loop0
sd $8,-8($4)
.L0: beq $6,$0,.Lend
nop
.Loop: ld $3,8($5)
daddiu $4,$4,32
daddiu $6,$6,-4
dsrl $11,$10,$7
dsll $12,$3,$13
ld $10,16($5)
dsrl $14,$3,$7
or $8,$11,$12
sd $8,-32($4)
dsll $9,$10,$13
ld $3,24($5)
dsrl $11,$10,$7
or $8,$14,$9
sd $8,-24($4)
dsll $12,$3,$13
ld $10,32($5)
dsrl $14,$3,$7
or $8,$11,$12
sd $8,-16($4)
dsll $9,$10,$13
daddiu $5,$5,32
or $8,$14,$9
bgtz $6,.Loop
sd $8,-8($4)
.Lend: dsrl $8,$10,$7
j $31
sd $8,0($4)
.end mpihelp_rshift

125
mpi/mips3/mpih-sub1.S Normal file
View File

@ -0,0 +1,125 @@
/* mips3 sub_n -- Subtract two limb vectors of the same length > 0 and
* store difference in a third limb vector.
*
* Copyright (C) 1995, 1998, 1999, 2000,
* 2001 Free Software Foundation, Inc.
*
* This file is part of GnuPG.
*
* GnuPG is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* GnuPG is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA
*/
/*******************
* mpi_limb_t
* mpihelp_sub_n( mpi_ptr_t res_ptr, (r4)
* mpi_ptr_t s1_ptr, (r5)
* mpi_ptr_t s2_ptr, (r6)
* mpi_size_t size) (r7)
*/
.text
.align 2
.globl mpihelp_sub_n
.ent mpihelp_sub_n
mpihelp_sub_n:
.set noreorder
.set nomacro
ld $10,0($5)
ld $11,0($6)
daddiu $7,$7,-1
and $9,$7,4-1 # number of limbs in first loop
beq $9,$0,.L0 # if multiple of 4 limbs, skip first loop
move $2,$0
dsubu $7,$7,$9
.Loop0: daddiu $9,$9,-1
ld $12,8($5)
daddu $11,$11,$2
ld $13,8($6)
sltu $8,$11,$2
dsubu $11,$10,$11
sltu $2,$10,$11
sd $11,0($4)
or $2,$2,$8
daddiu $5,$5,8
daddiu $6,$6,8
move $10,$12
move $11,$13
bne $9,$0,.Loop0
daddiu $4,$4,8
.L0: beq $7,$0,.Lend
nop
.Loop: daddiu $7,$7,-4
ld $12,8($5)
daddu $11,$11,$2
ld $13,8($6)
sltu $8,$11,$2
dsubu $11,$10,$11
sltu $2,$10,$11
sd $11,0($4)
or $2,$2,$8
ld $10,16($5)
daddu $13,$13,$2
ld $11,16($6)
sltu $8,$13,$2
dsubu $13,$12,$13
sltu $2,$12,$13
sd $13,8($4)
or $2,$2,$8
ld $12,24($5)
daddu $11,$11,$2
ld $13,24($6)
sltu $8,$11,$2
dsubu $11,$10,$11
sltu $2,$10,$11
sd $11,16($4)
or $2,$2,$8
ld $10,32($5)
daddu $13,$13,$2
ld $11,32($6)
sltu $8,$13,$2
dsubu $13,$12,$13
sltu $2,$12,$13
sd $13,24($4)
or $2,$2,$8
daddiu $5,$5,32
daddiu $6,$6,32
bne $7,$0,.Loop
daddiu $4,$4,32
.Lend: daddu $11,$11,$2
sltu $8,$11,$2
dsubu $11,$10,$11
sltu $2,$10,$11
sd $11,0($4)
j $31
or $2,$2,$8
.end mpihelp_sub_n

247
mpi/mpi-add.c Normal file
View File

@ -0,0 +1,247 @@
/* mpi-add.c - MPI functions
* Copyright (C) 1998, 1999, 2000, 2001 Free Software Foundation, Inc.
* Copyright (C) 1994, 1996 Free Software Foundation, Inc.
*
* This file is part of GnuPG.
*
* GnuPG is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* GnuPG is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA
*
* Note: This code is heavily based on the GNU MP Library.
* Actually it's the same code with only minor changes in the
* way the data is stored; this is to support the abstraction
* of an optional secure memory allocation which may be used
* to avoid revealing of sensitive data due to paging etc.
* The GNU MP Library itself is published under the LGPL;
* however I decided to publish this code under the plain GPL.
*/
#include <config.h>
#include <stdio.h>
#include <stdlib.h>
#include "mpi-internal.h"
/****************
* Add the unsigned integer V to the mpi-integer U and store the
* result in W. U and V may be the same.
*/
void
mpi_add_ui(MPI w, MPI u, unsigned long v )
{
mpi_ptr_t wp, up;
mpi_size_t usize, wsize;
int usign, wsign;
usize = u->nlimbs;
usign = u->sign;
wsign = 0;
/* If not space for W (and possible carry), increase space. */
wsize = usize + 1;
if( w->alloced < wsize )
mpi_resize(w, wsize);
/* These must be after realloc (U may be the same as W). */
up = u->d;
wp = w->d;
if( !usize ) { /* simple */
wp[0] = v;
wsize = v? 1:0;
}
else if( !usign ) { /* mpi is not negative */
mpi_limb_t cy;
cy = mpihelp_add_1(wp, up, usize, v);
wp[usize] = cy;
wsize = usize + cy;
}
else { /* The signs are different. Need exact comparison to determine
* which operand to subtract from which. */
if( usize == 1 && up[0] < v ) {
wp[0] = v - up[0];
wsize = 1;
}
else {
mpihelp_sub_1(wp, up, usize, v);
/* Size can decrease with at most one limb. */
wsize = usize - (wp[usize-1]==0);
wsign = 1;
}
}
w->nlimbs = wsize;
w->sign = wsign;
}
void
mpi_add(MPI w, MPI u, MPI v)
{
mpi_ptr_t wp, up, vp;
mpi_size_t usize, vsize, wsize;
int usign, vsign, wsign;
if( u->nlimbs < v->nlimbs ) { /* Swap U and V. */
usize = v->nlimbs;
usign = v->sign;
vsize = u->nlimbs;
vsign = u->sign;
wsize = usize + 1;
RESIZE_IF_NEEDED(w, wsize);
/* These must be after realloc (u or v may be the same as w). */
up = v->d;
vp = u->d;
}
else {
usize = u->nlimbs;
usign = u->sign;
vsize = v->nlimbs;
vsign = v->sign;
wsize = usize + 1;
RESIZE_IF_NEEDED(w, wsize);
/* These must be after realloc (u or v may be the same as w). */
up = u->d;
vp = v->d;
}
wp = w->d;
wsign = 0;
if( !vsize ) { /* simple */
MPN_COPY(wp, up, usize );
wsize = usize;
wsign = usign;
}
else if( usign != vsign ) { /* different sign */
/* This test is right since USIZE >= VSIZE */
if( usize != vsize ) {
mpihelp_sub(wp, up, usize, vp, vsize);
wsize = usize;
MPN_NORMALIZE(wp, wsize);
wsign = usign;
}
else if( mpihelp_cmp(up, vp, usize) < 0 ) {
mpihelp_sub_n(wp, vp, up, usize);
wsize = usize;
MPN_NORMALIZE(wp, wsize);
if( !usign )
wsign = 1;
}
else {
mpihelp_sub_n(wp, up, vp, usize);
wsize = usize;
MPN_NORMALIZE(wp, wsize);
if( usign )
wsign = 1;
}
}
else { /* U and V have same sign. Add them. */
mpi_limb_t cy = mpihelp_add(wp, up, usize, vp, vsize);
wp[usize] = cy;
wsize = usize + cy;
if( usign )
wsign = 1;
}
w->nlimbs = wsize;
w->sign = wsign;
}
/****************
* Subtract the unsigned integer V from the mpi-integer U and store the
* result in W.
*/
void
mpi_sub_ui(MPI w, MPI u, unsigned long v )
{
mpi_ptr_t wp, up;
mpi_size_t usize, wsize;
int usign, wsign;
usize = u->nlimbs;
usign = u->sign;
wsign = 0;
/* If not space for W (and possible carry), increase space. */
wsize = usize + 1;
if( w->alloced < wsize )
mpi_resize(w, wsize);
/* These must be after realloc (U may be the same as W). */
up = u->d;
wp = w->d;
if( !usize ) { /* simple */
wp[0] = v;
wsize = v? 1:0;
wsign = 1;
}
else if( usign ) { /* mpi and v are negative */
mpi_limb_t cy;
cy = mpihelp_add_1(wp, up, usize, v);
wp[usize] = cy;
wsize = usize + cy;
}
else { /* The signs are different. Need exact comparison to determine
* which operand to subtract from which. */
if( usize == 1 && up[0] < v ) {
wp[0] = v - up[0];
wsize = 1;
wsign = 1;
}
else {
mpihelp_sub_1(wp, up, usize, v);
/* Size can decrease with at most one limb. */
wsize = usize - (wp[usize-1]==0);
}
}
w->nlimbs = wsize;
w->sign = wsign;
}
void
mpi_sub(MPI w, MPI u, MPI v)
{
if( w == v ) {
MPI vv = mpi_copy(v);
vv->sign = !vv->sign;
mpi_add( w, u, vv );
mpi_free(vv);
}
else {
/* fixme: this is not thread-save (we temp. modify v) */
v->sign = !v->sign;
mpi_add( w, u, v );
v->sign = !v->sign;
}
}
void
mpi_addm( MPI w, MPI u, MPI v, MPI m)
{
mpi_add(w, u, v);
mpi_fdiv_r( w, w, m );
}
void
mpi_subm( MPI w, MPI u, MPI v, MPI m)
{
mpi_sub(w, u, v);
mpi_fdiv_r( w, w, m );
}

262
mpi/mpi-bit.c Normal file
View File

@ -0,0 +1,262 @@
/* mpi-bit.c - MPI bit level fucntions
* Copyright (C) 1998, 1999 Free Software Foundation, Inc.
*
* This file is part of GnuPG.
*
* GnuPG is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* GnuPG is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA
*/
#include <config.h>
#include <stdio.h>
#include <stdlib.h>
#include <assert.h>
#include "mpi-internal.h"
#include "longlong.h"
#ifdef MPI_INTERNAL_NEED_CLZ_TAB
#ifdef __STDC__
const
#endif
unsigned char
__clz_tab[] =
{
0,1,2,2,3,3,3,3,4,4,4,4,4,4,4,4,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,
6,6,6,6,6,6,6,6,6,6,6,6,6,6,6,6,6,6,6,6,6,6,6,6,6,6,6,6,6,6,6,6,
7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,
7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,
8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,
8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,
8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,
8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,
};
#endif
#define A_LIMB_1 ((mpi_limb_t)1)
/****************
* Sometimes we have MSL (most significant limbs) which are 0;
* this is for some reasons not good, so this function removes them.
*/
void
mpi_normalize( MPI a )
{
if( mpi_is_protected(a) )
return;
for( ; a->nlimbs && !a->d[a->nlimbs-1]; a->nlimbs-- )
;
}
/****************
* Return the number of bits in A.
*/
unsigned
mpi_get_nbits( MPI a )
{
unsigned n;
if( mpi_is_protected(a) ) {
n = mpi_get_nbit_info(a);
if( !n )
n = a->nlimbs * BITS_PER_MPI_LIMB;
return n;
}
mpi_normalize( a );
if( a->nlimbs ) {
mpi_limb_t alimb = a->d[a->nlimbs-1];
if( alimb )
count_leading_zeros( n, alimb );
else
n = BITS_PER_MPI_LIMB;
n = BITS_PER_MPI_LIMB - n + (a->nlimbs-1) * BITS_PER_MPI_LIMB;
}
else
n = 0;
return n;
}
/****************
* Test whether bit N is set.
*/
int
mpi_test_bit( MPI a, unsigned n )
{
unsigned limbno, bitno;
mpi_limb_t limb;
limbno = n / BITS_PER_MPI_LIMB;
bitno = n % BITS_PER_MPI_LIMB;
if( limbno >= a->nlimbs )
return 0; /* too far left: this is a 0 */
limb = a->d[limbno];
return (limb & (A_LIMB_1 << bitno))? 1: 0;
}
/****************
* Set bit N of A.
*/
void
mpi_set_bit( MPI a, unsigned n )
{
unsigned limbno, bitno;
limbno = n / BITS_PER_MPI_LIMB;
bitno = n % BITS_PER_MPI_LIMB;
if( limbno >= a->nlimbs ) { /* resize */
if( a->alloced >= limbno )
mpi_resize(a, limbno+1 );
a->nlimbs = limbno+1;
}
a->d[limbno] |= (A_LIMB_1<<bitno);
}
/****************
* Set bit N of A. and clear all bits above
*/
void
mpi_set_highbit( MPI a, unsigned n )
{
unsigned limbno, bitno;
limbno = n / BITS_PER_MPI_LIMB;
bitno = n % BITS_PER_MPI_LIMB;
if( limbno >= a->nlimbs ) { /* resize */
if( a->alloced >= limbno )
mpi_resize(a, limbno+1 );
a->nlimbs = limbno+1;
}
a->d[limbno] |= (A_LIMB_1<<bitno);
for( bitno++; bitno < BITS_PER_MPI_LIMB; bitno++ )
a->d[limbno] &= ~(A_LIMB_1 << bitno);
a->nlimbs = limbno+1;
}
/****************
* clear bit N of A and all bits above
*/
void
mpi_clear_highbit( MPI a, unsigned n )
{
unsigned limbno, bitno;
limbno = n / BITS_PER_MPI_LIMB;
bitno = n % BITS_PER_MPI_LIMB;
if( limbno >= a->nlimbs )
return; /* not allocated, so need to clear bits :-) */
for( ; bitno < BITS_PER_MPI_LIMB; bitno++ )
a->d[limbno] &= ~(A_LIMB_1 << bitno);
a->nlimbs = limbno+1;
}
/****************
* Clear bit N of A.
*/
void
mpi_clear_bit( MPI a, unsigned n )
{
unsigned limbno, bitno;
limbno = n / BITS_PER_MPI_LIMB;
bitno = n % BITS_PER_MPI_LIMB;
if( limbno >= a->nlimbs )
return; /* don't need to clear this bit, it's to far to left */
a->d[limbno] &= ~(A_LIMB_1 << bitno);
}
/****************
* Shift A by N bits to the right
* FIXME: should use alloc_limb if X and A are same.
*/
void
mpi_rshift( MPI x, MPI a, unsigned n )
{
mpi_ptr_t xp;
mpi_size_t xsize;
xsize = a->nlimbs;
x->sign = a->sign;
RESIZE_IF_NEEDED(x, xsize);
xp = x->d;
if( xsize ) {
mpihelp_rshift( xp, a->d, xsize, n);
MPN_NORMALIZE( xp, xsize);
}
x->nlimbs = xsize;
}
/****************
* Shift A by COUNT limbs to the left
* This is used only within the MPI library
*/
void
mpi_lshift_limbs( MPI a, unsigned int count )
{
mpi_ptr_t ap = a->d;
int n = a->nlimbs;
int i;
if( !count || !n )
return;
RESIZE_IF_NEEDED( a, n+count );
for( i = n-1; i >= 0; i-- )
ap[i+count] = ap[i];
for(i=0; i < count; i++ )
ap[i] = 0;
a->nlimbs += count;
}
/****************
* Shift A by COUNT limbs to the right
* This is used only within the MPI library
*/
void
mpi_rshift_limbs( MPI a, unsigned int count )
{
mpi_ptr_t ap = a->d;
mpi_size_t n = a->nlimbs;
unsigned int i;
if( count >= n ) {
a->nlimbs = 0;
return;
}
for( i = 0; i < n - count; i++ )
ap[i] = ap[i+count];
ap[i] = 0;
a->nlimbs -= count;
}

74
mpi/mpi-cmp.c Normal file
View File

@ -0,0 +1,74 @@
/* mpi-cmp.c - MPI functions
* Copyright (C) 1998, 1999 Free Software Foundation, Inc.
*
* This file is part of GnuPG.
*
* GnuPG is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* GnuPG is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA
*/
#include <config.h>
#include <stdio.h>
#include <stdlib.h>
#include "mpi-internal.h"
int
mpi_cmp_ui( MPI u, unsigned long v )
{
mpi_limb_t limb = v;
mpi_normalize( u );
if( !u->nlimbs && !limb )
return 0;
if( u->sign )
return -1;
if( u->nlimbs > 1 )
return 1;
if( u->d[0] == limb )
return 0;
else if( u->d[0] > limb )
return 1;
else
return -1;
}
int
mpi_cmp( MPI u, MPI v )
{
mpi_size_t usize, vsize;
int cmp;
mpi_normalize( u );
mpi_normalize( v );
usize = u->nlimbs;
vsize = v->nlimbs;
if( !u->sign && v->sign )
return 1;
if( u->sign && !v->sign )
return -1;
if( usize != vsize && !u->sign && !v->sign )
return usize - vsize;
if( usize != vsize && u->sign && v->sign )
return vsize + usize;
if( !usize )
return 0;
if( !(cmp=mpihelp_cmp( u->d, v->d, usize )) )
return 0;
if( (cmp < 0?1:0) == (u->sign?1:0))
return 1;
return -1;
}

322
mpi/mpi-div.c Normal file
View File

@ -0,0 +1,322 @@
/* mpi-div.c - MPI functions
* Copyright (C) 1994, 1996 Free Software Foundation, Inc.
* Copyright (C) 1998, 1999, 2000, 2001 Free Software Foundation, Inc.
*
* This file is part of GnuPG.
*
* GnuPG is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* GnuPG is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA
*
* Note: This code is heavily based on the GNU MP Library.
* Actually it's the same code with only minor changes in the
* way the data is stored; this is to support the abstraction
* of an optional secure memory allocation which may be used
* to avoid revealing of sensitive data due to paging etc.
* The GNU MP Library itself is published under the LGPL;
* however I decided to publish this code under the plain GPL.
*/
#include <config.h>
#include <stdio.h>
#include <stdlib.h>
#include "mpi-internal.h"
#include "longlong.h"
void
mpi_fdiv_r( MPI rem, MPI dividend, MPI divisor )
{
int divisor_sign = divisor->sign;
MPI temp_divisor = NULL;
/* We need the original value of the divisor after the remainder has been
* preliminary calculated. We have to copy it to temporary space if it's
* the same variable as REM. */
if( rem == divisor ) {
temp_divisor = mpi_copy( divisor );
divisor = temp_divisor;
}
mpi_tdiv_r( rem, dividend, divisor );
if( ((divisor_sign?1:0) ^ (dividend->sign?1:0)) && rem->nlimbs )
mpi_add( rem, rem, divisor);
if( temp_divisor )
mpi_free(temp_divisor);
}
/****************
* Division rounding the quotient towards -infinity.
* The remainder gets the same sign as the denominator.
* rem is optional
*/
ulong
mpi_fdiv_r_ui( MPI rem, MPI dividend, ulong divisor )
{
mpi_limb_t rlimb;
rlimb = mpihelp_mod_1( dividend->d, dividend->nlimbs, divisor );
if( rlimb && dividend->sign )
rlimb = divisor - rlimb;
if( rem ) {
rem->d[0] = rlimb;
rem->nlimbs = rlimb? 1:0;
}
return rlimb;
}
void
mpi_fdiv_q( MPI quot, MPI dividend, MPI divisor )
{
MPI tmp = mpi_alloc( mpi_get_nlimbs(quot) );
mpi_fdiv_qr( quot, tmp, dividend, divisor);
mpi_free(tmp);
}
void
mpi_fdiv_qr( MPI quot, MPI rem, MPI dividend, MPI divisor )
{
int divisor_sign = divisor->sign;
MPI temp_divisor = NULL;
if( quot == divisor || rem == divisor ) {
temp_divisor = mpi_copy( divisor );
divisor = temp_divisor;
}
mpi_tdiv_qr( quot, rem, dividend, divisor );
if( (divisor_sign ^ dividend->sign) && rem->nlimbs ) {
mpi_sub_ui( quot, quot, 1 );
mpi_add( rem, rem, divisor);
}
if( temp_divisor )
mpi_free(temp_divisor);
}
/* If den == quot, den needs temporary storage.
* If den == rem, den needs temporary storage.
* If num == quot, num needs temporary storage.
* If den has temporary storage, it can be normalized while being copied,
* i.e no extra storage should be allocated.
*/
void
mpi_tdiv_r( MPI rem, MPI num, MPI den)
{
mpi_tdiv_qr(NULL, rem, num, den );
}
void
mpi_tdiv_qr( MPI quot, MPI rem, MPI num, MPI den)
{
mpi_ptr_t np, dp;
mpi_ptr_t qp, rp;
mpi_size_t nsize = num->nlimbs;
mpi_size_t dsize = den->nlimbs;
mpi_size_t qsize, rsize;
mpi_size_t sign_remainder = num->sign;
mpi_size_t sign_quotient = num->sign ^ den->sign;
unsigned normalization_steps;
mpi_limb_t q_limb;
mpi_ptr_t marker[5];
int markidx=0;
/* Ensure space is enough for quotient and remainder.
* We need space for an extra limb in the remainder, because it's
* up-shifted (normalized) below. */
rsize = nsize + 1;
mpi_resize( rem, rsize);
qsize = rsize - dsize; /* qsize cannot be bigger than this. */
if( qsize <= 0 ) {
if( num != rem ) {
rem->nlimbs = num->nlimbs;
rem->sign = num->sign;
MPN_COPY(rem->d, num->d, nsize);
}
if( quot ) {
/* This needs to follow the assignment to rem, in case the
* numerator and quotient are the same. */
quot->nlimbs = 0;
quot->sign = 0;
}
return;
}
if( quot )
mpi_resize( quot, qsize);
/* Read pointers here, when reallocation is finished. */
np = num->d;
dp = den->d;
rp = rem->d;
/* Optimize division by a single-limb divisor. */
if( dsize == 1 ) {
mpi_limb_t rlimb;
if( quot ) {
qp = quot->d;
rlimb = mpihelp_divmod_1( qp, np, nsize, dp[0] );
qsize -= qp[qsize - 1] == 0;
quot->nlimbs = qsize;
quot->sign = sign_quotient;
}
else
rlimb = mpihelp_mod_1( np, nsize, dp[0] );
rp[0] = rlimb;
rsize = rlimb != 0?1:0;
rem->nlimbs = rsize;
rem->sign = sign_remainder;
return;
}
if( quot ) {
qp = quot->d;
/* Make sure QP and NP point to different objects. Otherwise the
* numerator would be gradually overwritten by the quotient limbs. */
if(qp == np) { /* Copy NP object to temporary space. */
np = marker[markidx++] = mpi_alloc_limb_space(nsize,
mpi_is_secure(quot));
MPN_COPY(np, qp, nsize);
}
}
else /* Put quotient at top of remainder. */
qp = rp + dsize;
count_leading_zeros( normalization_steps, dp[dsize - 1] );
/* Normalize the denominator, i.e. make its most significant bit set by
* shifting it NORMALIZATION_STEPS bits to the left. Also shift the
* numerator the same number of steps (to keep the quotient the same!).
*/
if( normalization_steps ) {
mpi_ptr_t tp;
mpi_limb_t nlimb;
/* Shift up the denominator setting the most significant bit of
* the most significant word. Use temporary storage not to clobber
* the original contents of the denominator. */
tp = marker[markidx++] = mpi_alloc_limb_space(dsize,mpi_is_secure(den));
mpihelp_lshift( tp, dp, dsize, normalization_steps );
dp = tp;
/* Shift up the numerator, possibly introducing a new most
* significant word. Move the shifted numerator in the remainder
* meanwhile. */
nlimb = mpihelp_lshift(rp, np, nsize, normalization_steps);
if( nlimb ) {
rp[nsize] = nlimb;
rsize = nsize + 1;
}
else
rsize = nsize;
}
else {
/* The denominator is already normalized, as required. Copy it to
* temporary space if it overlaps with the quotient or remainder. */
if( dp == rp || (quot && (dp == qp))) {
mpi_ptr_t tp;
tp = marker[markidx++] = mpi_alloc_limb_space(dsize, mpi_is_secure(den));
MPN_COPY( tp, dp, dsize );
dp = tp;
}
/* Move the numerator to the remainder. */
if( rp != np )
MPN_COPY(rp, np, nsize);
rsize = nsize;
}
q_limb = mpihelp_divrem( qp, 0, rp, rsize, dp, dsize );
if( quot ) {
qsize = rsize - dsize;
if(q_limb) {
qp[qsize] = q_limb;
qsize += 1;
}
quot->nlimbs = qsize;
quot->sign = sign_quotient;
}
rsize = dsize;
MPN_NORMALIZE (rp, rsize);
if( normalization_steps && rsize ) {
mpihelp_rshift(rp, rp, rsize, normalization_steps);
rsize -= rp[rsize - 1] == 0?1:0;
}
rem->nlimbs = rsize;
rem->sign = sign_remainder;
while( markidx )
mpi_free_limb_space(marker[--markidx]);
}
void
mpi_tdiv_q_2exp( MPI w, MPI u, unsigned count )
{
mpi_size_t usize, wsize;
mpi_size_t limb_cnt;
usize = u->nlimbs;
limb_cnt = count / BITS_PER_MPI_LIMB;
wsize = usize - limb_cnt;
if( limb_cnt >= usize )
w->nlimbs = 0;
else {
mpi_ptr_t wp;
mpi_ptr_t up;
RESIZE_IF_NEEDED( w, wsize );
wp = w->d;
up = u->d;
count %= BITS_PER_MPI_LIMB;
if( count ) {
mpihelp_rshift( wp, up + limb_cnt, wsize, count );
wsize -= !wp[wsize - 1];
}
else {
MPN_COPY_INCR( wp, up + limb_cnt, wsize);
}
w->nlimbs = wsize;
}
}
/****************
* Check whether dividend is divisible by divisor
* (note: divisor must fit into a limb)
*/
int
mpi_divisible_ui(MPI dividend, ulong divisor )
{
return !mpihelp_mod_1( dividend->d, dividend->nlimbs, divisor );
}

54
mpi/mpi-gcd.c Normal file
View File

@ -0,0 +1,54 @@
/* mpi-gcd.c - MPI functions
* Copyright (C) 1998, 1999, 2000, 2001 Free Software Foundation, Inc.
*
* This file is part of GnuPG.
*
* GnuPG is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* GnuPG is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA
*/
#include <config.h>
#include <stdio.h>
#include <stdlib.h>
#include "mpi-internal.h"
/****************
* Find the greatest common divisor G of A and B.
* Return: true if this 1, false in all other cases
*/
int
mpi_gcd( MPI g, MPI xa, MPI xb )
{
MPI a, b;
a = mpi_copy(xa);
b = mpi_copy(xb);
/* TAOCP Vol II, 4.5.2, Algorithm A */
a->sign = 0;
b->sign = 0;
while( mpi_cmp_ui( b, 0 ) ) {
mpi_fdiv_r( g, a, b ); /* g used as temorary variable */
mpi_set(a,b);
mpi_set(b,g);
}
mpi_set(g, a);
mpi_free(a);
mpi_free(b);
return !mpi_cmp_ui( g, 1);
}

36
mpi/mpi-inline.c Normal file
View File

@ -0,0 +1,36 @@
/* mpi-inline.c
* Copyright (C) 1999, 2000, 2001 Free Software Foundation, Inc.
*
* This file is part of GnuPG.
*
* GnuPG is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* GnuPG is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA
*/
#include <config.h>
#include <stdio.h>
#include <stdlib.h>
/* put the inline functions as real functions into the lib */
#define G10_MPI_INLINE_DECL
#include "mpi-internal.h"
/* always include the header becuase it is only
* included by mpi-internal if __GCC__ is defined but we
* need it here in all cases and the above definition of
* of the macro allows us to do so
*/
#include "mpi-inline.h"

Some files were not shown because too many files have changed in this diff Show More