beecrypt-3.0.0 merge: bring in latest asm code.

CVS patchset: 6783
CVS date: 2003/04/28 15:06:30

34 files changed, 3221 insertions, 3455 deletions

diff --git a/beecrypt/gas/Makefile.am b/beecrypt/gas/Makefile.am
index 8b80be914..4bcbc7cd5 100644
--- a/beecrypt/gas/Makefile.am
+++ b/beecrypt/gas/Makefile.am
@@ -1,9 +1,9 @@
 #
 # Makefile.am's purpose is to add the GNU Assembler sources to the dist
 #
-# Copyright (c) 2001, 2002 Virtual Unlimited B.V.
+# Copyright (c) 2001, 2002, 2003 Virtual Unlimited B.V.
 #
-# Author: Bob Deblier <bob@virtualunlimited.com>
+# Author: Bob Deblier <bob.deblier@pandora.be>
 #
 # This library is free software; you can redistribute it and/or
 # modify it under the terms of the GNU Lesser General Public
@@ -22,4 +22,4 @@
 
 AUTOMAKE_OPTIONS = gnu no-dependencies
 
-EXTRA_DIST = aesopt.i586.S aesopt.powerpc.S blowfishopt.i586.S blowfishopt.powerpc.S mp32opt.arm.S mp32opt.i386.S mp32opt.ia64.S mp32opt.powerpc.S mp32opt.sparcv8.S mp32opt.sparcv9.S mp64opt.ia64.S sha1opt.i586.S sha1opt.ia64.S sha1opt.powerpc.S
+EXTRA_DIST = aesopt.i586.m4 aesopt.ppc.m4 alpha.m4 asmdefs.m4 blowfishopt.i586.m4 blowfishopt.ppc.m4 ia64.m4 mpopt.alpha.m4 mpopt.arm.m4 mpopt.ia64.m4 mpopt.m68k.m4 mpopt.ppc.m4 mpopt.ppc64.m4 mpopt.sparcv8.m4 mpopt.sparcv8plus.m4 mpopt.x86.m4 ppc.m4 ppc64.m4 sha1opt.i586.m4 sparc.m4 x86.m4
diff --git a/beecrypt/gas/aesopt.i586.S b/beecrypt/gas/aesopt.i586.S
deleted file mode 100644
index deb0853b7..000000000
--- a/beecrypt/gas/aesopt.i586.S
+++ /dev/null
@@ -1,688 +0,0 @@
-/*
- * aesopt.i586.asm
- *
- * Assembler optimized AES routines for Intel Pentium processors
- *
- * Compile target is GNU Assembler
- *
- * Copyright (c) 2002 Bob Deblier <bob@virtualunlimited.com>
- *
- * This library is free software; you can redistribute it and/or
- * modify it under the terms of the GNU Lesser General Public
- * License as published by the Free Software Foundation; either
- * version 2.1 of the License, or (at your option) any later version.
- *
- * This library 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
- * Lesser General Public License for more details.
- *
- * You should have received a copy of the GNU Lesser General Public
- * License along with this library; if not, write to the Free Software
- * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
- *
- */
-
-#include "beecrypt.gas.h"
-
-	.file "aesopt.i586.S"
-
-	.text
-
-	.macro	sxrk
-	movl   (%esi),%eax
-	movl  4(%esi),%ebx
-	movl  8(%esi),%ecx
-	movl 12(%esi),%edx
-	bswap %eax
-	bswap %ebx
-	bswap %ecx
-	bswap %edx
-	xorl   (%ebp),%eax
-	xorl  4(%ebp),%ebx
-	xorl  8(%ebp),%ecx
-	xorl 12(%ebp),%edx
-	movl %eax,  (%esp)
-	movl %ebx, 4(%esp)
-	movl %ecx, 8(%esp)
-	movl %edx,12(%esp)
-	.endm
-
-	.macro	etfs	offset
-	movl \offset+ 0(%ebp),%ecx
-	movl \offset+ 4(%ebp),%edx
-
-	movzbl  3(%esp),%eax
-	movzbl  7(%esp),%ebx
-	xorl SYMBOL_NAME(_ae0)(,%eax,4),%ecx
-	xorl SYMBOL_NAME(_ae0)(,%ebx,4),%edx
-
-	movzbl  6(%esp),%eax
-	movzbl 10(%esp),%ebx
-	xorl SYMBOL_NAME(_ae1)(,%eax,4),%ecx
-	xorl SYMBOL_NAME(_ae1)(,%ebx,4),%edx
-
-	movzbl  9(%esp),%eax
-	movzbl 13(%esp),%ebx
-	xorl SYMBOL_NAME(_ae2)(,%eax,4),%ecx
-	xorl SYMBOL_NAME(_ae2)(,%ebx,4),%edx
-
-	movzbl 12(%esp),%eax
-	movzbl   (%esp),%ebx
-	xorl SYMBOL_NAME(_ae3)(,%eax,4),%ecx
-	xorl SYMBOL_NAME(_ae3)(,%ebx,4),%edx
-
-	movl %ecx,16(%esp)
-	movl %edx,20(%esp)
-
-	movl \offset+ 8(%ebp),%ecx
-	movl \offset+12(%ebp),%edx
-
-	movzbl 11(%esp),%eax
-	movzbl 15(%esp),%ebx
-	xorl SYMBOL_NAME(_ae0)(,%eax,4),%ecx
-	xorl SYMBOL_NAME(_ae0)(,%ebx,4),%edx
-
-	movzbl 14(%esp),%eax
-	movzbl  2(%esp),%ebx
-	xorl SYMBOL_NAME(_ae1)(,%eax,4),%ecx
-	xorl SYMBOL_NAME(_ae1)(,%ebx,4),%edx
-
-	movzbl  1(%esp),%eax
-	movzbl  5(%esp),%ebx
-	xorl SYMBOL_NAME(_ae2)(,%eax,4),%ecx
-	xorl SYMBOL_NAME(_ae2)(,%ebx,4),%edx
-
-	movzbl  4(%esp),%eax
-	movzbl  8(%esp),%ebx
-	xorl SYMBOL_NAME(_ae3)(,%eax,4),%ecx
-	xorl SYMBOL_NAME(_ae3)(,%ebx,4),%edx
-
-	movl %ecx,24(%esp)
-	movl %edx,28(%esp)
-	.endm
-
-	.macro	esft	offset
-	movl \offset+ 0(%ebp),%ecx
-	movl \offset+ 4(%ebp),%edx
-
-	movzbl 19(%esp),%eax
-	movzbl 23(%esp),%ebx
-	xorl SYMBOL_NAME(_ae0)(,%eax,4),%ecx
-	xorl SYMBOL_NAME(_ae0)(,%ebx,4),%edx
-
-	movzbl 22(%esp),%eax
-	movzbl 26(%esp),%ebx
-	xorl SYMBOL_NAME(_ae1)(,%eax,4),%ecx
-	xorl SYMBOL_NAME(_ae1)(,%ebx,4),%edx
-
-	movzbl 25(%esp),%eax
-	movzbl 29(%esp),%ebx
-	xorl SYMBOL_NAME(_ae2)(,%eax,4),%ecx
-	xorl SYMBOL_NAME(_ae2)(,%ebx,4),%edx
-
-	movzbl 28(%esp),%eax
-	movzbl 16(%esp),%ebx
-	xorl SYMBOL_NAME(_ae3)(,%eax,4),%ecx
-	xorl SYMBOL_NAME(_ae3)(,%ebx,4),%edx
-
-	movl %ecx,  (%esp)
-	movl %edx, 4(%esp)
-
-	movl \offset+ 8(%ebp),%ecx
-	movl \offset+12(%ebp),%edx
-
-	movzbl 27(%esp),%eax
-	movzbl 31(%esp),%ebx
-	xorl SYMBOL_NAME(_ae0)(,%eax,4),%ecx
-	xorl SYMBOL_NAME(_ae0)(,%ebx,4),%edx
-
-	movzbl 30(%esp),%eax
-	movzbl 18(%esp),%ebx
-	xorl SYMBOL_NAME(_ae1)(,%eax,4),%ecx
-	xorl SYMBOL_NAME(_ae1)(,%ebx,4),%edx
-
-	movzbl 17(%esp),%eax
-	movzbl 21(%esp),%ebx
-	xorl SYMBOL_NAME(_ae2)(,%eax,4),%ecx
-	xorl SYMBOL_NAME(_ae2)(,%ebx,4),%edx
-
-	movzbl 20(%esp),%eax
-	movzbl 24(%esp),%ebx
-	xorl SYMBOL_NAME(_ae3)(,%eax,4),%ecx
-	xorl SYMBOL_NAME(_ae3)(,%ebx,4),%edx
-
-	movl %ecx, 8(%esp)
-	movl %edx,12(%esp)
-	.endm
-
-	.macro	elr
-	movl 0(%ebp),%ecx
-	movl 4(%ebp),%edx
-
-	movzbl 19(%esp),%eax
-	movzbl 23(%esp),%ebx
-	movl SYMBOL_NAME(_ae4)(,%eax,4),%eax
-	movl SYMBOL_NAME(_ae4)(,%ebx,4),%ebx
-	andl $0xff000000,%eax
-	andl $0xff000000,%ebx
-	xorl %eax,%ecx
-	xorl %ebx,%edx
-
-	movzbl 22(%esp),%eax
-	movzbl 26(%esp),%ebx
-	movl SYMBOL_NAME(_ae4)(,%eax,4),%eax
-	movl SYMBOL_NAME(_ae4)(,%ebx,4),%ebx
-	andl $0xff0000,%eax
-	andl $0xff0000,%ebx
-	xorl %eax,%ecx
-	xorl %ebx,%edx
-
-	movzbl 25(%esp),%eax
-	movzbl 29(%esp),%ebx
-	movl SYMBOL_NAME(_ae4)(,%eax,4),%eax
-	movl SYMBOL_NAME(_ae4)(,%ebx,4),%ebx
-	andl $0xff00,%eax
-	andl $0xff00,%ebx
-	xorl %eax,%ecx
-	xorl %ebx,%edx
-
-	movzbl 28(%esp),%eax
-	movzbl 16(%esp),%ebx
-	movl SYMBOL_NAME(_ae4)(,%eax,4),%eax
-	movl SYMBOL_NAME(_ae4)(,%ebx,4),%ebx
-	andl $0xff,%eax
-	andl $0xff,%ebx
-	xorl %eax,%ecx
-	xorl %ebx,%edx
-
-	movl %ecx,  (%esp)
-	movl %edx, 4(%esp)
-
-	movl  8(%ebp),%ecx
-	movl 12(%ebp),%edx
-
-	movzbl 27(%esp),%eax
-	movzbl 31(%esp),%ebx
-	movl SYMBOL_NAME(_ae4)(,%eax,4),%eax
-	movl SYMBOL_NAME(_ae4)(,%ebx,4),%ebx
-	andl $0xff000000,%eax
-	andl $0xff000000,%ebx
-	xorl %eax,%ecx
-	xorl %ebx,%edx
-
-	movzbl 30(%esp),%eax
-	movzbl 18(%esp),%ebx
-	movl SYMBOL_NAME(_ae4)(,%eax,4),%eax
-	movl SYMBOL_NAME(_ae4)(,%ebx,4),%ebx
-	andl $0xff0000,%eax
-	andl $0xff0000,%ebx
-	xorl %eax,%ecx
-	xorl %ebx,%edx
-
-	movzbl 17(%esp),%eax
-	movzbl 21(%esp),%ebx
-	movl SYMBOL_NAME(_ae4)(,%eax,4),%eax
-	movl SYMBOL_NAME(_ae4)(,%ebx,4),%ebx
-	andl $0xff00,%eax
-	andl $0xff00,%ebx
-	xorl %eax,%ecx
-	xorl %ebx,%edx
-
-	movzbl 20(%esp),%eax
-	movzbl 24(%esp),%ebx
-	movl SYMBOL_NAME(_ae4)(,%eax,4),%eax
-	movl SYMBOL_NAME(_ae4)(,%ebx,4),%ebx
-	andl $0xff,%eax
-	andl $0xff,%ebx
-	xorl %eax,%ecx
-	xorl %ebx,%edx
-
-	movl %ecx, 8(%esp)
-	movl %edx,12(%esp)
-	.endm
-
-	.macro	eblock	label
-	sxrk
-
-	etfs 16
-	esft 32
-	etfs 48
-	esft 64
-	etfs 80
-	esft 96
-	etfs 112
-	esft 128
-	etfs 144
-
-	movl 256(%ebp),%eax
-	cmp $10,%eax
-	je \label
-
-	esft 160
-	etfs 176
-
-	movl 256(%ebp),%eax
-	cmp $12,%eax
-	je \label
-
-	esft 192
-	etfs 208
-
-	movl 256(%ebp),%eax
-
-	.align 4
-\label:
-	sall $4,%eax
-	addl %eax,%ebp
-
-	elr
-	.endm
-
-	.macro	dtfs	offset
-	movl \offset+0(%ebp),%ecx
-	movl \offset+4(%ebp),%edx
-
-	movzbl  3(%esp),%eax
-	movzbl  7(%esp),%ebx
-	xorl SYMBOL_NAME(_ad0)(,%eax,4),%ecx
-	xorl SYMBOL_NAME(_ad0)(,%ebx,4),%edx
-
-	movzbl 14(%esp),%eax
-	movzbl  2(%esp),%ebx
-	xorl SYMBOL_NAME(_ad1)(,%eax,4),%ecx
-	xorl SYMBOL_NAME(_ad1)(,%ebx,4),%edx
-
-	movzbl  9(%esp),%eax
-	movzbl 13(%esp),%ebx
-	xorl SYMBOL_NAME(_ad2)(,%eax,4),%ecx
-	xorl SYMBOL_NAME(_ad2)(,%ebx,4),%edx
-
-	movzbl  4(%esp),%eax
-	movzbl  8(%esp),%ebx
-	xorl SYMBOL_NAME(_ad3)(,%eax,4),%ecx
-	xorl SYMBOL_NAME(_ad3)(,%ebx,4),%edx
-
-	movl %ecx,16(%esp)
-	movl %edx,20(%esp)
-
-	movl \offset+ 8(%ebp),%ecx
-	movl \offset+12(%ebp),%edx
-
-	movzbl 11(%esp),%eax
-	movzbl 15(%esp),%ebx
-	xorl SYMBOL_NAME(_ad0)(,%eax,4),%ecx
-	xorl SYMBOL_NAME(_ad0)(,%ebx,4),%edx
-
-	movzbl  6(%esp),%eax
-	movzbl 10(%esp),%ebx
-	xorl SYMBOL_NAME(_ad1)(,%eax,4),%ecx
-	xorl SYMBOL_NAME(_ad1)(,%ebx,4),%edx
-
-	movzbl  1(%esp),%eax
-	movzbl  5(%esp),%ebx
-	xorl SYMBOL_NAME(_ad2)(,%eax,4),%ecx
-	xorl SYMBOL_NAME(_ad2)(,%ebx,4),%edx
-
-	movzbl 12(%esp),%eax
-	movzbl   (%esp),%ebx
-	xorl SYMBOL_NAME(_ad3)(,%eax,4),%ecx
-	xorl SYMBOL_NAME(_ad3)(,%ebx,4),%edx
-
-	movl %ecx,24(%esp)
-	movl %edx,28(%esp)
-	.endm
-
-	.macro	dsft	offset
-	movl \offset+ 0(%ebp),%ecx
-	movl \offset+ 4(%ebp),%edx
-
-	movzbl 19(%esp),%eax
-	movzbl 23(%esp),%ebx
-	xorl SYMBOL_NAME(_ad0)(,%eax,4),%ecx
-	xorl SYMBOL_NAME(_ad0)(,%ebx,4),%edx
-
-	movzbl 30(%esp),%eax
-	movzbl 18(%esp),%ebx
-	xorl SYMBOL_NAME(_ad1)(,%eax,4),%ecx
-	xorl SYMBOL_NAME(_ad1)(,%ebx,4),%edx
-
-	movzbl 25(%esp),%eax
-	movzbl 29(%esp),%ebx
-	xorl SYMBOL_NAME(_ad2)(,%eax,4),%ecx
-	xorl SYMBOL_NAME(_ad2)(,%ebx,4),%edx
-
-	movzbl 20(%esp),%eax
-	movzbl 24(%esp),%ebx
-	xorl SYMBOL_NAME(_ad3)(,%eax,4),%ecx
-	xorl SYMBOL_NAME(_ad3)(,%ebx,4),%edx
-
-	movl %ecx,  (%esp)
-	movl %edx, 4(%esp)
-
-	movl \offset+ 8(%ebp),%ecx
-	movl \offset+12(%ebp),%edx
-
-	movzbl 27(%esp),%eax
-	movzbl 31(%esp),%ebx
-	xorl SYMBOL_NAME(_ad0)(,%eax,4),%ecx
-	xorl SYMBOL_NAME(_ad0)(,%ebx,4),%edx
-
-	movzbl 22(%esp),%eax
-	movzbl 26(%esp),%ebx
-	xorl SYMBOL_NAME(_ad1)(,%eax,4),%ecx
-	xorl SYMBOL_NAME(_ad1)(,%ebx,4),%edx
-
-	movzbl 17(%esp),%eax
-	movzbl 21(%esp),%ebx
-	xorl SYMBOL_NAME(_ad2)(,%eax,4),%ecx
-	xorl SYMBOL_NAME(_ad2)(,%ebx,4),%edx
-
-	movzbl 28(%esp),%eax
-	movzbl 16(%esp),%ebx
-	xorl SYMBOL_NAME(_ad3)(,%eax,4),%ecx
-	xorl SYMBOL_NAME(_ad3)(,%ebx,4),%edx
-
-	movl %ecx, 8(%esp)
-	movl %edx,12(%esp)
-	.endm
-
-	.macro	dlr
-	movl 0(%ebp),%ecx
-	movl 4(%ebp),%edx
-
-	movzbl 19(%esp),%eax
-	movzbl 23(%esp),%ebx
-	movl SYMBOL_NAME(_ad4)(,%eax,4),%eax
-	movl SYMBOL_NAME(_ad4)(,%ebx,4),%ebx
-	andl $0xff000000,%eax
-	andl $0xff000000,%ebx
-	xorl %eax,%ecx
-	xorl %ebx,%edx
-
-	movzbl 30(%esp),%eax
-	movzbl 18(%esp),%ebx
-	movl SYMBOL_NAME(_ad4)(,%eax,4),%eax
-	movl SYMBOL_NAME(_ad4)(,%ebx,4),%ebx
-	andl $0xff0000,%eax
-	andl $0xff0000,%ebx
-	xorl %eax,%ecx
-	xorl %ebx,%edx
-
-	movzbl 25(%esp),%eax
-	movzbl 29(%esp),%ebx
-	movl SYMBOL_NAME(_ad4)(,%eax,4),%eax
-	movl SYMBOL_NAME(_ad4)(,%ebx,4),%ebx
-	andl $0xff00,%eax
-	andl $0xff00,%ebx
-	xorl %eax,%ecx
-	xorl %ebx,%edx
-
-	movzbl 20(%esp),%eax
-	movzbl 24(%esp),%ebx
-	movl SYMBOL_NAME(_ad4)(,%eax,4),%eax
-	movl SYMBOL_NAME(_ad4)(,%ebx,4),%ebx
-	andl $0xff,%eax
-	andl $0xff,%ebx
-	xorl %eax,%ecx
-	xorl %ebx,%edx
-
-	movl %ecx,  (%esp)
-	movl %edx, 4(%esp)
-
-	movl  8(%ebp),%ecx
-	movl 12(%ebp),%edx
-
-	movzbl 27(%esp),%eax
-	movzbl 31(%esp),%ebx
-	movl SYMBOL_NAME(_ad4)(,%eax,4),%eax
-	movl SYMBOL_NAME(_ad4)(,%ebx,4),%ebx
-	andl $0xff000000,%eax
-	andl $0xff000000,%ebx
-	xorl %eax,%ecx
-	xorl %ebx,%edx
-
-	movzbl 22(%esp),%eax
-	movzbl 26(%esp),%ebx
-	movl SYMBOL_NAME(_ad4)(,%eax,4),%eax
-	movl SYMBOL_NAME(_ad4)(,%ebx,4),%ebx
-	andl $0xff0000,%eax
-	andl $0xff0000,%ebx
-	xorl %eax,%ecx
-	xorl %ebx,%edx
-
-	movzbl 17(%esp),%eax
-	movzbl 21(%esp),%ebx
-	movl SYMBOL_NAME(_ad4)(,%eax,4),%eax
-	movl SYMBOL_NAME(_ad4)(,%ebx,4),%ebx
-	andl $0xff00,%eax
-	andl $0xff00,%ebx
-	xorl %eax,%ecx
-	xorl %ebx,%edx
-
-	movzbl 28(%esp),%eax
-	movzbl 16(%esp),%ebx
-	movl SYMBOL_NAME(_ad4)(,%eax,4),%eax
-	movl SYMBOL_NAME(_ad4)(,%ebx,4),%ebx
-	andl $0xff,%eax
-	andl $0xff,%ebx
-	xorl %eax,%ecx
-	xorl %ebx,%edx
-
-	movl %ecx, 8(%esp)
-	movl %edx,12(%esp)
-	.endm
-
-	.macro	dblock	label
-	sxrk
-
-	dtfs 16
-	dsft 32
-	dtfs 48
-	dsft 64
-	dtfs 80
-	dsft 96
-	dtfs 112
-	dsft 128
-	dtfs 144
-
-	movl 256(%ebp),%eax
-	cmp $10,%eax
-	je \label
-
-	dsft 160
-	dtfs 176
-
-	movl 256(%ebp),%eax
-	cmp $12,%eax
-	je \label
-
-	dsft 192
-	dtfs 208
-
-	movl 256(%ebp),%eax
-
-	.align 4
-\label:
-	sall $4,%eax
-	addl %eax,%ebp
-
-	dlr
-	.endm
-
-C_FUNCTION_BEGIN(aesEncrypt)
-LABEL(aesEncrypt)
-	pushl %edi
-	pushl %esi
-	pushl %ebp
-	pushl %ebx
-
-	movl 20(%esp),%ebp
-	movl 24(%esp),%edi
-	movl 28(%esp),%esi
-
-	subl $32,%esp
-
-	eblock LOCAL(00)
-
-	movl   (%esp),%eax
-	movl  4(%esp),%ebx
-	movl  8(%esp),%ecx
-	movl 12(%esp),%edx
-	bswap %eax
-	bswap %ebx
-	bswap %ecx
-	bswap %edx
-	movl %eax,  (%edi)
-	movl %ebx, 4(%edi)
-	movl %ecx, 8(%edi)
-	movl %edx,12(%edi)
-
-	addl $32,%esp
-
-	xorl %eax,%eax
-
-	popl %ebx
-	popl %ebp
-	popl %esi
-	popl %edi
-	ret
-C_FUNCTION_END(aesEncrypt, LOCAL(aesEncrypt_size))
-
-
-C_FUNCTION_BEGIN(aesDecrypt)
-LABEL(aesDecrypt)
-	pushl %edi
-	pushl %esi
-	pushl %ebp
-	pushl %ebx
-
-	movl 20(%esp),%ebp
-	movl 24(%esp),%edi
-	movl 28(%esp),%esi
-
-	subl $32,%esp
-
-	dblock LOCAL(01)
-
-	movl   (%esp),%eax
-	movl  4(%esp),%ebx
-	movl  8(%esp),%ecx
-	movl 12(%esp),%edx
-	bswap %eax
-	bswap %ebx
-	bswap %ecx
-	bswap %edx
-	movl %eax,  (%edi)
-	movl %ebx, 4(%edi)
-	movl %ecx, 8(%edi)
-	movl %edx,12(%edi)
-
-	addl $32,%esp
-
-	xorl %eax,%eax
-
-	popl %ebx
-	popl %ebp
-	popl %esi
-	popl %edi
-	ret
-C_FUNCTION_END(aesDecrypt, LOCAL(aesDecrypt_size))
-
-
-C_FUNCTION_BEGIN(aesECBEncrypt)
-LABEL(aesECBEncrypt)
-	pushl %edi
-	pushl %esi
-	pushl %ebp
-	pushl %ebx
-
-	movl 28(%esp),%edi
-	movl 32(%esp),%esi
-
-	subl $32,%esp
-
-	.align 4
-LOCAL(02):
-	movl 52(%esp),%ebp
-
-	eblock LOCAL(03)
-
-	movl   (%esp),%eax
-	movl  4(%esp),%ebx
-	movl  8(%esp),%ecx
-	movl 12(%esp),%edx
-	bswap %eax
-	bswap %ebx
-	bswap %ecx
-	bswap %edx
-	movl %eax,  (%edi)
-	movl %ebx, 4(%edi)
-	movl %ecx, 8(%edi)
-	movl %edx,12(%edi)
-
-	addl $16,%esi
-	addl $16,%edi
-
-	decl 56(%esp)
-	jnz LOCAL(02)
-
-	addl $32,%esp
-
-	xorl %eax,%eax
-
-	popl %ebx
-	popl %ebp
-	popl %esi
-	popl %edi
-	ret
-C_FUNCTION_END(aesECBEncrypt, LOCAL(aesECBEncrypt_size))
-
-
-C_FUNCTION_BEGIN(aesECBDecrypt)
-LABEL(aesECBDecrypt)
-	pushl %edi
-	pushl %esi
-	pushl %ebp
-	pushl %ebx
-
-	movl 28(%esp),%edi
-	movl 32(%esp),%esi
-
-	subl $32,%esp
-
-	.align 4
-LOCAL(04):
-	movl 52(%esp),%ebp
-
-	dblock LOCAL(05)
-
-	movl   (%esp),%eax
-	movl  4(%esp),%ebx
-	movl  8(%esp),%ecx
-	movl 12(%esp),%edx
-	bswap %eax
-	bswap %ebx
-	bswap %ecx
-	bswap %edx
-	movl %eax,  (%edi)
-	movl %ebx, 4(%edi)
-	movl %ecx, 8(%edi)
-	movl %edx,12(%edi)
-
-	addl $16,%esi
-	addl $16,%edi
-
-	decl 56(%esp)
-	jnz LOCAL(04)
-
-	addl $32,%esp
-
-	xorl %eax,%eax
-
-	popl %ebx
-	popl %ebp
-	popl %esi
-	popl %edi
-	ret
-C_FUNCTION_END(aesECBDecrypt, LOCAL(aesECBDecrypt_size))
diff --git a/beecrypt/gas/aesopt.i586.m4 b/beecrypt/gas/aesopt.i586.m4
new file mode 100644
index 000000000..e8dbf1da5
--- /dev/null
+++ b/beecrypt/gas/aesopt.i586.m4
@@ -0,0 +1,580 @@
+dnl  aesopt.i586.m4
+dnl
+dnl  Copyright (c) 2003 Bob Deblier
+dnl 
+dnl  Author: Bob Deblier <bob.deblier@pandora.be>
+dnl 
+dnl  This library is free software; you can redistribute it and/or
+dnl  modify it under the terms of the GNU Lesser General Public
+dnl  License as published by the Free Software Foundation; either
+dnl  version 2.1 of the License, or (at your option) any later version.
+dnl 
+dnl  This library 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 GNU
+dnl  Lesser General Public License for more details.
+dnl 
+dnl  You should have received a copy of the GNU Lesser General Public
+dnl  License along with this library; if not, write to the Free Software
+dnl  Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
+
+include(config.m4)
+include(ASM_SRCDIR/x86.m4)
+
+define(`sxrk',`
+	movl   (%esi),%eax
+	movl  4(%esi),%ebx
+	movl  8(%esi),%ecx
+	movl 12(%esi),%edx
+	bswap %eax
+	bswap %ebx
+	bswap %ecx
+	bswap %edx
+	xorl   (%ebp),%eax
+	xorl  4(%ebp),%ebx
+	xorl  8(%ebp),%ecx
+	xorl 12(%ebp),%edx
+	movl %eax,  (%esp)
+	movl %ebx, 4(%esp)
+	movl %ecx, 8(%esp)
+	movl %edx,12(%esp)
+')
+
+define(`etfs',`
+	movl $1+0(%ebp),%ecx
+	movl $1+4(%ebp),%edx
+
+	movzbl  3(%esp),%eax
+	movzbl  7(%esp),%ebx
+	xorl SYMNAME(_ae0)(,%eax,4),%ecx
+	xorl SYMNAME(_ae0)(,%ebx,4),%edx
+
+	movzbl  6(%esp),%eax
+	movzbl 10(%esp),%ebx
+	xorl SYMNAME(_ae1)(,%eax,4),%ecx
+	xorl SYMNAME(_ae1)(,%ebx,4),%edx
+
+	movzbl  9(%esp),%eax
+	movzbl 13(%esp),%ebx
+	xorl SYMNAME(_ae2)(,%eax,4),%ecx
+	xorl SYMNAME(_ae2)(,%ebx,4),%edx
+
+	movzbl 12(%esp),%eax
+	movzbl   (%esp),%ebx
+	xorl SYMNAME(_ae3)(,%eax,4),%ecx
+	xorl SYMNAME(_ae3)(,%ebx,4),%edx
+
+	movl %ecx,16(%esp)
+	movl %edx,20(%esp)
+
+	movl $1+ 8(%ebp),%ecx
+	movl $1+12(%ebp),%edx
+
+	movzbl 11(%esp),%eax
+	movzbl 15(%esp),%ebx
+	xorl SYMNAME(_ae0)(,%eax,4),%ecx
+	xorl SYMNAME(_ae0)(,%ebx,4),%edx
+
+	movzbl 14(%esp),%eax
+	movzbl  2(%esp),%ebx
+	xorl SYMNAME(_ae1)(,%eax,4),%ecx
+	xorl SYMNAME(_ae1)(,%ebx,4),%edx
+
+	movzbl  1(%esp),%eax
+	movzbl  5(%esp),%ebx
+	xorl SYMNAME(_ae2)(,%eax,4),%ecx
+	xorl SYMNAME(_ae2)(,%ebx,4),%edx
+
+	movzbl  4(%esp),%eax
+	movzbl  8(%esp),%ebx
+	xorl SYMNAME(_ae3)(,%eax,4),%ecx
+	xorl SYMNAME(_ae3)(,%ebx,4),%edx
+
+	movl %ecx,24(%esp)
+	movl %edx,28(%esp)
+')
+
+define(`esft',`
+	movl $1+0(%ebp),%ecx
+	movl $1+4(%ebp),%edx
+
+	movzbl 19(%esp),%eax
+	movzbl 23(%esp),%ebx
+	xorl SYMNAME(_ae0)(,%eax,4),%ecx
+	xorl SYMNAME(_ae0)(,%ebx,4),%edx
+
+	movzbl 22(%esp),%eax
+	movzbl 26(%esp),%ebx
+	xorl SYMNAME(_ae1)(,%eax,4),%ecx
+	xorl SYMNAME(_ae1)(,%ebx,4),%edx
+
+	movzbl 25(%esp),%eax
+	movzbl 29(%esp),%ebx
+	xorl SYMNAME(_ae2)(,%eax,4),%ecx
+	xorl SYMNAME(_ae2)(,%ebx,4),%edx
+
+	movzbl 28(%esp),%eax
+	movzbl 16(%esp),%ebx
+	xorl SYMNAME(_ae3)(,%eax,4),%ecx
+	xorl SYMNAME(_ae3)(,%ebx,4),%edx
+
+	movl %ecx,  (%esp)
+	movl %edx, 4(%esp)
+
+	movl $1+ 8(%ebp),%ecx
+	movl $1+12(%ebp),%edx
+
+	movzbl 27(%esp),%eax
+	movzbl 31(%esp),%ebx
+	xorl SYMNAME(_ae0)(,%eax,4),%ecx
+	xorl SYMNAME(_ae0)(,%ebx,4),%edx
+
+	movzbl 30(%esp),%eax
+	movzbl 18(%esp),%ebx
+	xorl SYMNAME(_ae1)(,%eax,4),%ecx
+	xorl SYMNAME(_ae1)(,%ebx,4),%edx
+
+	movzbl 17(%esp),%eax
+	movzbl 21(%esp),%ebx
+	xorl SYMNAME(_ae2)(,%eax,4),%ecx
+	xorl SYMNAME(_ae2)(,%ebx,4),%edx
+
+	movzbl 20(%esp),%eax
+	movzbl 24(%esp),%ebx
+	xorl SYMNAME(_ae3)(,%eax,4),%ecx
+	xorl SYMNAME(_ae3)(,%ebx,4),%edx
+
+	movl %ecx, 8(%esp)
+	movl %edx,12(%esp)
+')
+
+define(`elr',`
+	movl 0(%ebp),%ecx
+	movl 4(%ebp),%edx
+
+	movzbl 19(%esp),%eax
+	movzbl 23(%esp),%ebx
+	movl SYMNAME(_ae4)(,%eax,4),%eax
+	movl SYMNAME(_ae4)(,%ebx,4),%ebx
+	andl `$'0xff000000,%eax
+	andl `$'0xff000000,%ebx
+	xorl %eax,%ecx
+	xorl %ebx,%edx
+
+	movzbl 22(%esp),%eax
+	movzbl 26(%esp),%ebx
+	movl SYMNAME(_ae4)(,%eax,4),%eax
+	movl SYMNAME(_ae4)(,%ebx,4),%ebx
+	andl `$'0xff0000,%eax
+	andl `$'0xff0000,%ebx
+	xorl %eax,%ecx
+	xorl %ebx,%edx
+
+	movzbl 25(%esp),%eax
+	movzbl 29(%esp),%ebx
+	movl SYMNAME(_ae4)(,%eax,4),%eax
+	movl SYMNAME(_ae4)(,%ebx,4),%ebx
+	andl `$'0xff00,%eax
+	andl `$'0xff00,%ebx
+	xorl %eax,%ecx
+	xorl %ebx,%edx
+
+	movzbl 28(%esp),%eax
+	movzbl 16(%esp),%ebx
+	movl SYMNAME(_ae4)(,%eax,4),%eax
+	movl SYMNAME(_ae4)(,%ebx,4),%ebx
+	andl `$'0xff,%eax
+	andl `$'0xff,%ebx
+	xorl %eax,%ecx
+	xorl %ebx,%edx
+
+	movl %ecx,  (%esp)
+	movl %edx, 4(%esp)
+
+	movl  8(%ebp),%ecx
+	movl 12(%ebp),%edx
+
+	movzbl 27(%esp),%eax
+	movzbl 31(%esp),%ebx
+	movl SYMNAME(_ae4)(,%eax,4),%eax
+	movl SYMNAME(_ae4)(,%ebx,4),%ebx
+	andl `$'0xff000000,%eax
+	andl `$'0xff000000,%ebx
+	xorl %eax,%ecx
+	xorl %ebx,%edx
+
+	movzbl 30(%esp),%eax
+	movzbl 18(%esp),%ebx
+	movl SYMNAME(_ae4)(,%eax,4),%eax
+	movl SYMNAME(_ae4)(,%ebx,4),%ebx
+	andl `$'0xff0000,%eax
+	andl `$'0xff0000,%ebx
+	xorl %eax,%ecx
+	xorl %ebx,%edx
+
+	movzbl 17(%esp),%eax
+	movzbl 21(%esp),%ebx
+	movl SYMNAME(_ae4)(,%eax,4),%eax
+	movl SYMNAME(_ae4)(,%ebx,4),%ebx
+	andl `$'0xff00,%eax
+	andl `$'0xff00,%ebx
+	xorl %eax,%ecx
+	xorl %ebx,%edx
+
+	movzbl 20(%esp),%eax
+	movzbl 24(%esp),%ebx
+	movl SYMNAME(_ae4)(,%eax,4),%eax
+	movl SYMNAME(_ae4)(,%ebx,4),%ebx
+	andl `$'0xff,%eax
+	andl `$'0xff,%ebx
+	xorl %eax,%ecx
+	xorl %ebx,%edx
+
+	movl %ecx, 8(%esp)
+	movl %edx,12(%esp)
+')
+
+define(`eblock',`
+	sxrk
+
+	etfs(16)
+	esft(32)
+	etfs(48)
+	esft(64)
+	etfs(80)
+	esft(96)
+	etfs(112)
+	esft(128)
+	etfs(144)
+
+	movl 256(%ebp),%eax
+	cmp `$'10,%eax
+	je $1
+
+	esft(160)
+	etfs(176)
+
+	movl 256(%ebp),%eax
+	cmp `$'12,%eax
+	je $1
+
+	esft(192)
+	etfs(208)
+
+	movl 256(%ebp),%eax
+
+	.align 4
+$1:
+	sall `$'4,%eax
+	addl %eax,%ebp
+
+	elr
+')
+
+define(`dtfs',`
+	movl $1+0(%ebp),%ecx
+	movl $1+4(%ebp),%edx
+
+	movzbl  3(%esp),%eax
+	movzbl  7(%esp),%ebx
+	xorl SYMNAME(_ad0)(,%eax,4),%ecx
+	xorl SYMNAME(_ad0)(,%ebx,4),%edx
+
+	movzbl 14(%esp),%eax
+	movzbl  2(%esp),%ebx
+	xorl SYMNAME(_ad1)(,%eax,4),%ecx
+	xorl SYMNAME(_ad1)(,%ebx,4),%edx
+
+	movzbl  9(%esp),%eax
+	movzbl 13(%esp),%ebx
+	xorl SYMNAME(_ad2)(,%eax,4),%ecx
+	xorl SYMNAME(_ad2)(,%ebx,4),%edx
+
+	movzbl  4(%esp),%eax
+	movzbl  8(%esp),%ebx
+	xorl SYMNAME(_ad3)(,%eax,4),%ecx
+	xorl SYMNAME(_ad3)(,%ebx,4),%edx
+
+	movl %ecx,16(%esp)
+	movl %edx,20(%esp)
+
+	movl $1+ 8(%ebp),%ecx
+	movl $1+12(%ebp),%edx
+
+	movzbl 11(%esp),%eax
+	movzbl 15(%esp),%ebx
+	xorl SYMNAME(_ad0)(,%eax,4),%ecx
+	xorl SYMNAME(_ad0)(,%ebx,4),%edx
+
+	movzbl  6(%esp),%eax
+	movzbl 10(%esp),%ebx
+	xorl SYMNAME(_ad1)(,%eax,4),%ecx
+	xorl SYMNAME(_ad1)(,%ebx,4),%edx
+
+	movzbl  1(%esp),%eax
+	movzbl  5(%esp),%ebx
+	xorl SYMNAME(_ad2)(,%eax,4),%ecx
+	xorl SYMNAME(_ad2)(,%ebx,4),%edx
+
+	movzbl 12(%esp),%eax
+	movzbl   (%esp),%ebx
+	xorl SYMNAME(_ad3)(,%eax,4),%ecx
+	xorl SYMNAME(_ad3)(,%ebx,4),%edx
+
+	movl %ecx,24(%esp)
+	movl %edx,28(%esp)
+')
+
+define(`dsft',`
+	movl $1+0(%ebp),%ecx
+	movl $1+4(%ebp),%edx
+
+	movzbl 19(%esp),%eax
+	movzbl 23(%esp),%ebx
+	xorl SYMNAME(_ad0)(,%eax,4),%ecx
+	xorl SYMNAME(_ad0)(,%ebx,4),%edx
+
+	movzbl 30(%esp),%eax
+	movzbl 18(%esp),%ebx
+	xorl SYMNAME(_ad1)(,%eax,4),%ecx
+	xorl SYMNAME(_ad1)(,%ebx,4),%edx
+
+	movzbl 25(%esp),%eax
+	movzbl 29(%esp),%ebx
+	xorl SYMNAME(_ad2)(,%eax,4),%ecx
+	xorl SYMNAME(_ad2)(,%ebx,4),%edx
+
+	movzbl 20(%esp),%eax
+	movzbl 24(%esp),%ebx
+	xorl SYMNAME(_ad3)(,%eax,4),%ecx
+	xorl SYMNAME(_ad3)(,%ebx,4),%edx
+
+	movl %ecx,  (%esp)
+	movl %edx, 4(%esp)
+
+	movl $1+ 8(%ebp),%ecx
+	movl $1+12(%ebp),%edx
+
+	movzbl 27(%esp),%eax
+	movzbl 31(%esp),%ebx
+	xorl SYMNAME(_ad0)(,%eax,4),%ecx
+	xorl SYMNAME(_ad0)(,%ebx,4),%edx
+
+	movzbl 22(%esp),%eax
+	movzbl 26(%esp),%ebx
+	xorl SYMNAME(_ad1)(,%eax,4),%ecx
+	xorl SYMNAME(_ad1)(,%ebx,4),%edx
+
+	movzbl 17(%esp),%eax
+	movzbl 21(%esp),%ebx
+	xorl SYMNAME(_ad2)(,%eax,4),%ecx
+	xorl SYMNAME(_ad2)(,%ebx,4),%edx
+
+	movzbl 28(%esp),%eax
+	movzbl 16(%esp),%ebx
+	xorl SYMNAME(_ad3)(,%eax,4),%ecx
+	xorl SYMNAME(_ad3)(,%ebx,4),%edx
+
+	movl %ecx, 8(%esp)
+	movl %edx,12(%esp)
+')
+
+define(`dlr',`
+	movl 0(%ebp),%ecx
+	movl 4(%ebp),%edx
+
+	movzbl 19(%esp),%eax
+	movzbl 23(%esp),%ebx
+	movl SYMNAME(_ad4)(,%eax,4),%eax
+	movl SYMNAME(_ad4)(,%ebx,4),%ebx
+	andl `$'0xff000000,%eax
+	andl `$'0xff000000,%ebx
+	xorl %eax,%ecx
+	xorl %ebx,%edx
+
+	movzbl 30(%esp),%eax
+	movzbl 18(%esp),%ebx
+	movl SYMNAME(_ad4)(,%eax,4),%eax
+	movl SYMNAME(_ad4)(,%ebx,4),%ebx
+	andl `$'0xff0000,%eax
+	andl `$'0xff0000,%ebx
+	xorl %eax,%ecx
+	xorl %ebx,%edx
+
+	movzbl 25(%esp),%eax
+	movzbl 29(%esp),%ebx
+	movl SYMNAME(_ad4)(,%eax,4),%eax
+	movl SYMNAME(_ad4)(,%ebx,4),%ebx
+	andl `$'0xff00,%eax
+	andl `$'0xff00,%ebx
+	xorl %eax,%ecx
+	xorl %ebx,%edx
+
+	movzbl 20(%esp),%eax
+	movzbl 24(%esp),%ebx
+	movl SYMNAME(_ad4)(,%eax,4),%eax
+	movl SYMNAME(_ad4)(,%ebx,4),%ebx
+	andl `$'0xff,%eax
+	andl `$'0xff,%ebx
+	xorl %eax,%ecx
+	xorl %ebx,%edx
+
+	movl %ecx,  (%esp)
+	movl %edx, 4(%esp)
+
+	movl  8(%ebp),%ecx
+	movl 12(%ebp),%edx
+
+	movzbl 27(%esp),%eax
+	movzbl 31(%esp),%ebx
+	movl SYMNAME(_ad4)(,%eax,4),%eax
+	movl SYMNAME(_ad4)(,%ebx,4),%ebx
+	andl `$'0xff000000,%eax
+	andl `$'0xff000000,%ebx
+	xorl %eax,%ecx
+	xorl %ebx,%edx
+
+	movzbl 22(%esp),%eax
+	movzbl 26(%esp),%ebx
+	movl SYMNAME(_ad4)(,%eax,4),%eax
+	movl SYMNAME(_ad4)(,%ebx,4),%ebx
+	andl `$'0xff0000,%eax
+	andl `$'0xff0000,%ebx
+	xorl %eax,%ecx
+	xorl %ebx,%edx
+
+	movzbl 17(%esp),%eax
+	movzbl 21(%esp),%ebx
+	movl SYMNAME(_ad4)(,%eax,4),%eax
+	movl SYMNAME(_ad4)(,%ebx,4),%ebx
+	andl `$'0xff00,%eax
+	andl `$'0xff00,%ebx
+	xorl %eax,%ecx
+	xorl %ebx,%edx
+
+	movzbl 28(%esp),%eax
+	movzbl 16(%esp),%ebx
+	movl SYMNAME(_ad4)(,%eax,4),%eax
+	movl SYMNAME(_ad4)(,%ebx,4),%ebx
+	andl `$'0xff,%eax
+	andl `$'0xff,%ebx
+	xorl %eax,%ecx
+	xorl %ebx,%edx
+
+	movl %ecx, 8(%esp)
+	movl %edx,12(%esp)
+')
+
+define(`dblock',`
+	sxrk
+
+	dtfs(16)
+	dsft(32)
+	dtfs(48)
+	dsft(64)
+	dtfs(80)
+	dsft(96)
+	dtfs(112)
+	dsft(128)
+	dtfs(144)
+
+	movl 256(%ebp),%eax
+	cmp `$'10,%eax
+	je $1
+
+	dsft(160)
+	dtfs(176)
+
+	movl 256(%ebp),%eax
+	cmp `$'12,%eax
+	je $1
+
+	dsft(192)
+	dtfs(208)
+
+	movl 256(%ebp),%eax
+
+	.align 4
+$1:
+	sall `$'4,%eax
+	addl %eax,%ebp
+
+	dlr
+')
+
+C_FUNCTION_BEGIN(aesEncrypt)
+	pushl %edi
+	pushl %esi
+	pushl %ebp
+	pushl %ebx
+
+	movl 20(%esp),%ebp
+	movl 24(%esp),%edi
+	movl 28(%esp),%esi
+
+	subl `$'32,%esp
+
+	eblock(LOCAL(00))
+
+	movl   (%esp),%eax
+	movl  4(%esp),%ebx
+	movl  8(%esp),%ecx
+	movl 12(%esp),%edx
+	bswap %eax
+	bswap %ebx
+	bswap %ecx
+	bswap %edx
+	movl %eax,  (%edi)
+	movl %ebx, 4(%edi)
+	movl %ecx, 8(%edi)
+	movl %edx,12(%edi)
+
+	addl `$'32,%esp
+
+	xorl %eax,%eax
+
+	popl %ebx
+	popl %ebp
+	popl %esi
+	popl %edi
+	ret
+C_FUNCTION_END(aesEncrypt)
+
+
+C_FUNCTION_BEGIN(aesDecrypt)
+	pushl %edi
+	pushl %esi
+	pushl %ebp
+	pushl %ebx
+
+	movl 20(%esp),%ebp
+	movl 24(%esp),%edi
+	movl 28(%esp),%esi
+
+	subl `$'32,%esp
+
+	dblock(LOCAL(01))
+
+	movl   (%esp),%eax
+	movl  4(%esp),%ebx
+	movl  8(%esp),%ecx
+	movl 12(%esp),%edx
+	bswap %eax
+	bswap %ebx
+	bswap %ecx
+	bswap %edx
+	movl %eax,  (%edi)
+	movl %ebx, 4(%edi)
+	movl %ecx, 8(%edi)
+	movl %edx,12(%edi)
+
+	addl `$'32,%esp
+
+	xorl %eax,%eax
+
+	popl %ebx
+	popl %ebp
+	popl %esi
+	popl %edi
+	ret
+C_FUNCTION_END(aesDecrypt)
diff --git a/beecrypt/gas/aesopt.powerpc.S b/beecrypt/gas/aesopt.powerpc.S
deleted file mode 100644
index 17e36dc41..000000000
--- a/beecrypt/gas/aesopt.powerpc.S
+++ /dev/null
@@ -1,683 +0,0 @@
-/*
- * aesopt.powerpc.asm
- *
- * Assembler optimized AES routines for PowerPC processors
- *
- * Compile target is GNU Assembler
- *
- * Copyright (c) 2002 Bob Deblier <bob@virtualunlimited.com>
- *
- * This library is free software; you can redistribute it and/or
- * modify it under the terms of the GNU Lesser General Public
- * License as published by the Free Software Foundation; either
- * version 2.1 of the License, or (at your option) any later version.
- *
- * This library 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
- * Lesser General Public License for more details.
- *
- * You should have received a copy of the GNU Lesser General Public
- * License along with this library; if not, write to the Free Software
- * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
- *
- */
-
-#include "beecrypt.gas.h"
-
-	.file "aesopt.powerpc.S"
-
-	.text
-
-/*
- * ae0 in r11
- * r7 thru r10 scratch
- * four registers for s: r24 r25 r26 r27
- * four registers for t: r28 r29 r30 r31
- */
-
-#define s0 r24
-#define s1 r25
-#define s2 r26
-#define s3 r27
-#define t0 r28
-#define t1 r29
-#define t2 r30
-#define t3 r31
-
-	.macro	sxrk rk src
-	#if WORDS_BIGENDIAN
-	lwz s0, 0(\src)
-	lwz s1, 4(\src)
-	lwz s2, 8(\src)
-	lwz s3,12(\src)
-	lwz r7, 0(\rk)
-	lwz r8, 4(\rk)
-	lwz r9, 8(\rk)
-	lwz r10,12(\rk)
-	#else
-	# error ppc little-endian not implemented
-	#endif
-	xor s0,s0,r7
-	xor s1,s1,r8
-	xor s2,s2,r9
-	xor s3,s3,r10
-	.endm
-
-	.macro	etfs	rk offset
-	lwz t0,\offset+ 0(\rk)
-	lwz t1,\offset+ 4(\rk)
-	lwz t2,\offset+ 8(\rk)
-	lwz t3,\offset+12(\rk)
-
-	rlwinm r7,s0,10,22,29 /* ((s0 >> 24) & 0xff) * 4 */
-	rlwinm r8,s1,10,22,29 /* ((s1 >> 24) & 0xff) * 4 */
-	rlwinm r9,s2,10,22,29 /* ((s2 >> 24) & 0xff) * 4 */
-	rlwinm r10,s3,10,22,29 /* ((s3 >> 24) & 0xff) * 4 */
-	lwzx r7,r7,r12
-	lwzx r8,r8,r12
-	lwzx r9,r9,r12
-	lwzx r10,r10,r12
-	xor t0,t0,r7
-	xor t1,t1,r8
-	xor t2,t2,r9
-	xor t3,t3,r10
-
-	la r12,1024(r12)
-
-	rlwinm r7,s1,18,22,29 /* ((s1 >> 16) & 0xff) * 4 */
-	rlwinm r8,s2,18,22,29 /* ((s2 >> 16) & 0xff) * 4 */
-	rlwinm r9,s3,18,22,29 /* ((s3 >> 16) & 0xff) * 4 */
-	rlwinm r10,s0,18,22,29 /* ((s0 >> 16) & 0xff) * 4 */
-	lwzx r7,r7,r12
-	lwzx r8,r8,r12
-	lwzx r9,r9,r12
-	lwzx r10,r10,r12
-	xor t0,t0,r7
-	xor t1,t1,r8
-	xor t2,t2,r9
-	xor t3,t3,r10
-
-	la r12,1024(r12)
-
-	rlwinm r7,s2,26,22,29 /* ((s2 >> 8) & 0xff) * 4 */
-	rlwinm r8,s3,26,22,29 /* ((s3 >> 8) & 0xff) * 4 */
-	rlwinm r9,s0,26,22,29 /* ((s0 >> 8) & 0xff) * 4 */
-	rlwinm r10,s1,26,22,29 /* ((s1 >> 8) & 0xff) * 4 */
-	lwzx r7,r7,r12
-	lwzx r8,r8,r12
-	lwzx r9,r9,r12
-	lwzx r10,r10,r12
-	xor t0,t0,r7
-	xor t1,t1,r8
-	xor t2,t2,r9
-	xor t3,t3,r10
-
-	la r12,1024(r12)
-
-	rlwinm r7,s3,2,22,29 /* ((s3 >> 0) & 0xff) * 4 */
-	rlwinm r8,s0,2,22,29 /* ((s0 >> 0) & 0xff) * 4 */
-	rlwinm r9,s1,2,22,29 /* ((s1 >> 0) & 0xff) * 4 */
-	rlwinm r10,s2,2,22,29 /* ((s2 >> 0) & 0xff) * 4 */
-	lwzx r7,r7,r12
-	lwzx r8,r8,r12
-	lwzx r9,r9,r12
-	lwzx r10,r10,r12
-	xor t0,t0,r7
-	xor t1,t1,r8
-	xor t2,t2,r9
-	xor t3,t3,r10
-
-	la r12,-3072(r12)
-	.endm
-
-	.macro	esft	rk offset
-	lwz s0,\offset+ 0(\rk)
-	lwz s1,\offset+ 4(\rk)
-	lwz s2,\offset+ 8(\rk)
-	lwz s3,\offset+12(\rk)
-
-	rlwinm r7,t0,10,22,29 /* ((t0 >> 24) & 0xff) * 4 */
-	rlwinm r8,t1,10,22,29 /* ((t1 >> 24) & 0xff) * 4 */
-	rlwinm r9,t2,10,22,29 /* ((t2 >> 24) & 0xff) * 4 */
-	rlwinm r10,t3,10,22,29 /* ((s3 >> 24) & 0xff) * 4 */
-	lwzx r7,r7,r12
-	lwzx r8,r8,r12
-	lwzx r9,r9,r12
-	lwzx r10,r10,r12
-	xor s0,s0,r7
-	xor s1,s1,r8
-	xor s2,s2,r9
-	xor s3,s3,r10
-
-	la r12,1024(r12)
-
-	rlwinm r7,t1,18,22,29 /* ((t1 >> 16) & 0xff) * 4 */
-	rlwinm r8,t2,18,22,29 /* ((t2 >> 16) & 0xff) * 4 */
-	rlwinm r9,t3,18,22,29 /* ((t3 >> 16) & 0xff) * 4 */
-	rlwinm r10,t0,18,22,29 /* ((t0 >> 16) & 0xff) * 4 */
-	lwzx r7,r7,r12
-	lwzx r8,r8,r12
-	lwzx r9,r9,r12
-	lwzx r10,r10,r12
-	xor s0,s0,r7
-	xor s1,s1,r8
-	xor s2,s2,r9
-	xor s3,s3,r10
-
-	la r12,1024(r12)
-
-	rlwinm r7,t2,26,22,29 /* ((t2 >> 8) & 0xff) * 4 */
-	rlwinm r8,t3,26,22,29 /* ((t3 >> 8) & 0xff) * 4 */
-	rlwinm r9,t0,26,22,29 /* ((t0 >> 8) & 0xff) * 4 */
-	rlwinm r10,t1,26,22,29 /* ((t1 >> 8) & 0xff) * 4 */
-	lwzx r7,r7,r12
-	lwzx r8,r8,r12
-	lwzx r9,r9,r12
-	lwzx r10,r10,r12
-	xor s0,s0,r7
-	xor s1,s1,r8
-	xor s2,s2,r9
-	xor s3,s3,r10
-
-	la r12,1024(r12)
-
-	rlwinm r7,t3,2,22,29 /* ((t3 >> 0) & 0xff) * 4 */
-	rlwinm r8,t0,2,22,29 /* ((t0 >> 0) & 0xff) * 4 */
-	rlwinm r9,t1,2,22,29 /* ((t1 >> 0) & 0xff) * 4 */
-	rlwinm r10,t2,2,22,29 /* ((t2 >> 0) & 0xff) * 4 */
-	lwzx r7,r7,r12
-	lwzx r8,r8,r12
-	lwzx r9,r9,r12
-	lwzx r10,r10,r12
-	xor s0,s0,r7
-	xor s1,s1,r8
-	xor s2,s2,r9
-	xor s3,s3,r10
-
-	la r12,-3072(r12)
-	.endm
-
-	.macro	elr rk
-	lwz s0, 0(\rk)
-	lwz s1, 4(\rk)
-	lwz s2, 8(\rk)
-	lwz s3,12(\rk)
-
-	la r12,4096(r12)
-
-	rlwinm r7,t0,10,22,29 /* ((t0 >> 24) & 0xff) * 4 */
-	rlwinm r8,t1,10,22,29 /* ((t1 >> 24) & 0xff) * 4 */
-	rlwinm r9,t2,10,22,29 /* ((t2 >> 24) & 0xff) * 4 */
-	rlwinm r10,t3,10,22,29 /* ((t3 >> 24) & 0xff) * 4 */
-	lwzx r7,r7,r12
-	lwzx r8,r8,r12
-	lwzx r9,r9,r12
-	lwzx r10,r10,r12
-	rlwinm r7,r7,0,0,7 /* & 0xff000000 */
-	rlwinm r8,r8,0,0,7 /* & 0xff000000 */
-	rlwinm r9,r9,0,0,7 /* & 0xff000000 */
-	rlwinm r10,r10,0,0,7 /* & 0xff000000 */
-	xor s0,s0,r7
-	xor s1,s1,r8
-	xor s2,s2,r9
-	xor s3,s3,r10
-
-	rlwinm r7,t1,18,22,29 /* ((t1 >> 16) & 0xff) * 4 */
-	rlwinm r8,t2,18,22,29 /* ((t2 >> 16) & 0xff) * 4 */
-	rlwinm r9,t3,18,22,29 /* ((t3 >> 16) & 0xff) * 4 */
-	rlwinm r10,t0,18,22,29 /* ((t0 >> 16) & 0xff) * 4 */
-	lwzx r7,r7,r12
-	lwzx r8,r8,r12
-	lwzx r9,r9,r12
-	lwzx r10,r10,r12
-	rlwinm r7,r7,0,8,15 /* & 0xff0000 */
-	rlwinm r8,r8,0,8,15 /* & 0xff0000 */
-	rlwinm r9,r9,0,8,15 /* & 0xff0000 */
-	rlwinm r10,r10,0,8,15 /* & 0xff0000 */
-	xor s0,s0,r7
-	xor s1,s1,r8
-	xor s2,s2,r9
-	xor s3,s3,r10
-
-	rlwinm r7,t2,26,22,29 /* ((t2 >> 8) & 0xff) * 4 */
-	rlwinm r8,t3,26,22,29 /* ((t3 >> 8) & 0xff) * 4 */
-	rlwinm r9,t0,26,22,29 /* ((t0 >> 8) & 0xff) * 4 */
-	rlwinm r10,t1,26,22,29 /* ((t1 >> 8) & 0xff) * 4 */
-	lwzx r7,r7,r12
-	lwzx r8,r8,r12
-	lwzx r9,r9,r12
-	lwzx r10,r10,r12
-	rlwinm r7,r7,0,16,23 /* & 0xff00 */
-	rlwinm r8,r8,0,16,23 /* & 0xff00 */
-	rlwinm r9,r9,0,16,23 /* & 0xff00 */
-	rlwinm r10,r10,0,16,23 /* & 0xff00 */
-	xor s0,s0,r7
-	xor s1,s1,r8
-	xor s2,s2,r9
-	xor s3,s3,r10
-
-	rlwinm r7,t3,2,22,29 /* ((t3 >> 0) & 0xff) * 4 */
-	rlwinm r8,t0,2,22,29 /* ((t0 >> 0) & 0xff) * 4 */
-	rlwinm r9,t1,2,22,29 /* ((t1 >> 0) & 0xff) * 4 */
-	rlwinm r10,t2,2,22,29 /* ((t2 >> 0) & 0xff) * 4 */
-	lwzx r7,r7,r12
-	lwzx r8,r8,r12
-	lwzx r9,r9,r12
-	lwzx r10,r10,r12
-	rlwinm r7,r7,0,24,31 /* & 0xff */
-	rlwinm r8,r8,0,24,31 /* & 0xff */
-	rlwinm r9,r9,0,24,31 /* & 0xff */
-	rlwinm r10,r10,0,24,31 /* & 0xff */
-	xor s0,s0,r7
-	xor s1,s1,r8
-	xor s2,s2,r9
-	xor s3,s3,r10
-
-	la r12,-4096(r12)
-	.endm
-
-	.macro	eblock	rk src label
-	sxrk \rk \src
-
-	etfs \rk 16
-	esft \rk 32
-	etfs \rk 48
-	esft \rk 64
-	etfs \rk 80
-	esft \rk 96
-	etfs \rk 112
-	esft \rk 128
-	etfs \rk 144
-
-	lwz r11,256(\rk)
-	cmpwi r11,10
-	beq \label
-
-	esft \rk 160
-	etfs \rk 176
-
-	cmpwi r11,12
-	beq \label
-
-	esft \rk 192
-	etfs \rk 208
-
-\label:
-	slwi r11,r11,4
-	add \rk,\rk,r11
-
-	elr \rk
-	.endm
-
-	.macro	dtfs	rk offset
-	lwz t0,\offset+ 0(\rk)
-	lwz t1,\offset+ 4(\rk)
-	lwz t2,\offset+ 8(\rk)
-	lwz t3,\offset+12(\rk)
-
-	rlwinm r7,s0,10,22,29 /* ((s0 >> 24) & 0xff) * 4 */
-	rlwinm r8,s1,10,22,29 /* ((s1 >> 24) & 0xff) * 4 */
-	rlwinm r9,s2,10,22,29 /* ((s2 >> 24) & 0xff) * 4 */
-	rlwinm r10,s3,10,22,29 /* ((s3 >> 24) & 0xff) * 4 */
-	lwzx r7,r7,r12
-	lwzx r8,r8,r12
-	lwzx r9,r9,r12
-	lwzx r10,r10,r12
-	xor t0,t0,r7
-	xor t1,t1,r8
-	xor t2,t2,r9
-	xor t3,t3,r10
-
-	la r12,1024(r12)
-
-	rlwinm r7,s3,18,22,29 /* ((s3 >> 16) & 0xff) * 4 */
-	rlwinm r8,s0,18,22,29 /* ((s0 >> 16) & 0xff) * 4 */
-	rlwinm r9,s1,18,22,29 /* ((s1 >> 16) & 0xff) * 4 */
-	rlwinm r10,s2,18,22,29 /* ((s2 >> 16) & 0xff) * 4 */
-
-/* start here */
-	lwzx r7,r7,r12
-	lwzx r8,r8,r12
-	lwzx r9,r9,r12
-	lwzx r10,r10,r12
-	xor t0,t0,r7
-	xor t1,t1,r8
-	xor t2,t2,r9
-	xor t3,t3,r10
-
-	la r12,1024(r12)
-
-	rlwinm r7,s2,26,22,29 /* ((s2 >> 8) & 0xff) * 4 */
-	rlwinm r8,s3,26,22,29 /* ((s3 >> 8) & 0xff) * 4 */
-	rlwinm r9,s0,26,22,29 /* ((s0 >> 8) & 0xff) * 4 */
-	rlwinm r10,s1,26,22,29 /* ((s1 >> 8) & 0xff) * 4 */
-	lwzx r7,r7,r12
-	lwzx r8,r8,r12
-	lwzx r9,r9,r12
-	lwzx r10,r10,r12
-	xor t0,t0,r7
-	xor t1,t1,r8
-	xor t2,t2,r9
-	xor t3,t3,r10
-
-	la r12,1024(r12)
-
-	rlwinm r7,s1,2,22,29 /* ((s1 >> 0) & 0xff) * 4 */
-	rlwinm r8,s2,2,22,29 /* ((s2 >> 0) & 0xff) * 4 */
-	rlwinm r9,s3,2,22,29 /* ((s3 >> 0) & 0xff) * 4 */
-	rlwinm r10,s0,2,22,29 /* ((s0 >> 0) & 0xff) * 4 */
-	lwzx r7,r7,r12
-	lwzx r8,r8,r12
-	lwzx r9,r9,r12
-	lwzx r10,r10,r12
-	xor t0,t0,r7
-	xor t1,t1,r8
-	xor t2,t2,r9
-	xor t3,t3,r10
-
-	la r12,-3072(r12)
-	.endm
-
-	.macro	dsft	rk offset
-	lwz s0,\offset+ 0(\rk)
-	lwz s1,\offset+ 4(\rk)
-	lwz s2,\offset+ 8(\rk)
-	lwz s3,\offset+12(\rk)
-
-	rlwinm r7,t0,10,22,29 /* ((t0 >> 24) & 0xff) * 4 */
-	rlwinm r8,t1,10,22,29 /* ((t1 >> 24) & 0xff) * 4 */
-	rlwinm r9,t2,10,22,29 /* ((t2 >> 24) & 0xff) * 4 */
-	rlwinm r10,t3,10,22,29 /* ((s3 >> 24) & 0xff) * 4 */
-	lwzx r7,r7,r12
-	lwzx r8,r8,r12
-	lwzx r9,r9,r12
-	lwzx r10,r10,r12
-	xor s0,s0,r7
-	xor s1,s1,r8
-	xor s2,s2,r9
-	xor s3,s3,r10
-
-	la r12,1024(r12)
-
-	rlwinm r7,t3,18,22,29 /* ((t3 >> 16) & 0xff) * 4 */
-	rlwinm r8,t0,18,22,29 /* ((t0 >> 16) & 0xff) * 4 */
-	rlwinm r9,t1,18,22,29 /* ((t1 >> 16) & 0xff) * 4 */
-	rlwinm r10,t2,18,22,29 /* ((t2 >> 16) & 0xff) * 4 */
-	lwzx r7,r7,r12
-	lwzx r8,r8,r12
-	lwzx r9,r9,r12
-	lwzx r10,r10,r12
-	xor s0,s0,r7
-	xor s1,s1,r8
-	xor s2,s2,r9
-	xor s3,s3,r10
-
-	la r12,1024(r12)
-
-	rlwinm r7,t2,26,22,29 /* ((t2 >> 8) & 0xff) * 4 */
-	rlwinm r8,t3,26,22,29 /* ((t3 >> 8) & 0xff) * 4 */
-	rlwinm r9,t0,26,22,29 /* ((t0 >> 8) & 0xff) * 4 */
-	rlwinm r10,t1,26,22,29 /* ((t1 >> 8) & 0xff) * 4 */
-	lwzx r7,r7,r12
-	lwzx r8,r8,r12
-	lwzx r9,r9,r12
-	lwzx r10,r10,r12
-	xor s0,s0,r7
-	xor s1,s1,r8
-	xor s2,s2,r9
-	xor s3,s3,r10
-
-	la r12,1024(r12)
-
-	rlwinm r7,t1,2,22,29 /* ((t1 >> 0) & 0xff) * 4 */
-	rlwinm r8,t2,2,22,29 /* ((t2 >> 0) & 0xff) * 4 */
-	rlwinm r9,t3,2,22,29 /* ((t3 >> 0) & 0xff) * 4 */
-	rlwinm r10,t0,2,22,29 /* ((t0 >> 0) & 0xff) * 4 */
-	lwzx r7,r7,r12
-	lwzx r8,r8,r12
-	lwzx r9,r9,r12
-	lwzx r10,r10,r12
-	xor s0,s0,r7
-	xor s1,s1,r8
-	xor s2,s2,r9
-	xor s3,s3,r10
-
-	la r12,-3072(r12)
-	.endm
-
-	.macro	dlr rk
-	lwz s0, 0(\rk)
-	lwz s1, 4(\rk)
-	lwz s2, 8(\rk)
-	lwz s3,12(\rk)
-
-	la r12,4096(r12)
-
-	rlwinm r7,t0,10,22,29 /* ((t0 >> 24) & 0xff) * 4 */
-	rlwinm r8,t1,10,22,29 /* ((t1 >> 24) & 0xff) * 4 */
-	rlwinm r9,t2,10,22,29 /* ((t2 >> 24) & 0xff) * 4 */
-	rlwinm r10,t3,10,22,29 /* ((t3 >> 24) & 0xff) * 4 */
-	lwzx r7,r7,r12
-	lwzx r8,r8,r12
-	lwzx r9,r9,r12
-	lwzx r10,r10,r12
-	rlwinm r7,r7,0,0,7 /* & 0xff000000 */
-	rlwinm r8,r8,0,0,7 /* & 0xff000000 */
-	rlwinm r9,r9,0,0,7 /* & 0xff000000 */
-	rlwinm r10,r10,0,0,7 /* & 0xff000000 */
-	xor s0,s0,r7
-	xor s1,s1,r8
-	xor s2,s2,r9
-	xor s3,s3,r10
-
-	rlwinm r7,t3,18,22,29 /* ((t3 >> 16) & 0xff) * 4 */
-	rlwinm r8,t0,18,22,29 /* ((t0 >> 16) & 0xff) * 4 */
-	rlwinm r9,t1,18,22,29 /* ((t1 >> 16) & 0xff) * 4 */
-	rlwinm r10,t2,18,22,29 /* ((t2 >> 16) & 0xff) * 4 */
-	lwzx r7,r7,r12
-	lwzx r8,r8,r12
-	lwzx r9,r9,r12
-	lwzx r10,r10,r12
-	rlwinm r7,r7,0,8,15 /* & 0xff0000 */
-	rlwinm r8,r8,0,8,15 /* & 0xff0000 */
-	rlwinm r9,r9,0,8,15 /* & 0xff0000 */
-	rlwinm r10,r10,0,8,15 /* & 0xff0000 */
-	xor s0,s0,r7
-	xor s1,s1,r8
-	xor s2,s2,r9
-	xor s3,s3,r10
-
-	rlwinm r7,t2,26,22,29 /* ((t2 >> 8) & 0xff) * 4 */
-	rlwinm r8,t3,26,22,29 /* ((t3 >> 8) & 0xff) * 4 */
-	rlwinm r9,t0,26,22,29 /* ((t0 >> 8) & 0xff) * 4 */
-	rlwinm r10,t1,26,22,29 /* ((t1 >> 8) & 0xff) * 4 */
-	lwzx r7,r7,r12
-	lwzx r8,r8,r12
-	lwzx r9,r9,r12
-	lwzx r10,r10,r12
-	rlwinm r7,r7,0,16,23 /* & 0xff00 */
-	rlwinm r8,r8,0,16,23 /* & 0xff00 */
-	rlwinm r9,r9,0,16,23 /* & 0xff00 */
-	rlwinm r10,r10,0,16,23 /* & 0xff00 */
-	xor s0,s0,r7
-	xor s1,s1,r8
-	xor s2,s2,r9
-	xor s3,s3,r10
-
-	rlwinm r7,t1,2,22,29 /* ((t1 >> 0) & 0xff) * 4 */
-	rlwinm r8,t2,2,22,29 /* ((t2 >> 0) & 0xff) * 4 */
-	rlwinm r9,t3,2,22,29 /* ((t3 >> 0) & 0xff) * 4 */
-	rlwinm r10,t0,2,22,29 /* ((t0 >> 0) & 0xff) * 4 */
-	lwzx r7,r7,r12
-	lwzx r8,r8,r12
-	lwzx r9,r9,r12
-	lwzx r10,r10,r12
-	rlwinm r7,r7,0,24,31 /* & 0xff */
-	rlwinm r8,r8,0,24,31 /* & 0xff */
-	rlwinm r9,r9,0,24,31 /* & 0xff */
-	rlwinm r10,r10,0,24,31 /* & 0xff */
-	xor s0,s0,r7
-	xor s1,s1,r8
-	xor s2,s2,r9
-	xor s3,s3,r10
-
-	la r12,-4096(r12)
-	.endm
-
-	.macro	dblock	rk src label
-	sxrk \rk \src
-
-	dtfs \rk 16
-	dsft \rk 32
-	dtfs \rk 48
-	dsft \rk 64
-	dtfs \rk 80
-	dsft \rk 96
-	dtfs \rk 112
-	dsft \rk 128
-	dtfs \rk 144
-
-	lwz r11,256(\rk)
-	cmpwi r11,10
-	beq \label
-
-	dsft \rk 160
-	dtfs \rk 176
-
-	cmpwi r11,12
-	beq \label
-
-	dsft \rk 192
-	dtfs \rk 208
-
-\label:
-	slwi r11,r11,4
-	add \rk,\rk,r11
-
-	dlr \rk
-	.endm
-
-C_FUNCTION_BEGIN(aesEncrypt)
-LABEL(aesEncrypt)
-	subi r1,r1,32
-	stmw r24,0(r1)
-
-	lis r12,_ae0@ha
-	la r12,_ae0@l(r12)
-
-	eblock rk=r3 src=r5 label=LOCAL(00)
-
-	#if WORDS_BIGENDIAN
-	stw s0, 0(r4)
-	stw s1, 4(r4)
-	stw s2, 8(r4)
-	stw s3,12(r4)
-	#else
-	# error ppc little-endian mode not supported
-	#endif
-
-	li r3,0
-	lmw r24,0(r1)
-	addi r1,r1,32
-	blr
-C_FUNCTION_END(aesEncrypt, LOCAL(aesEncrypt_size))
-
-
-C_FUNCTION_BEGIN(aesDecrypt)
-LABEL(aesDecrypt)
-	subi r1,r1,32
-	stmw r24,0(r1)
-
-	lis r12,_ad0@ha
-	la r12,_ad0@l(r12)
-
-	dblock rk=r3 src=r5 label=LOCAL(01)
-
-	#if WORDS_BIGENDIAN
-	stw s0, 0(r4)
-	stw s1, 4(r4)
-	stw s2, 8(r4)
-	stw s3,12(r4)
-	#else
-	# error ppc little-endian mode not supported
-	#endif
-
-	li r3,0
-	lmw r24,0(r1)
-	addi r1,r1,32
-	blr
-C_FUNCTION_END(aesDecrypt, LOCAL(aesDecrypt_size))
-
-C_FUNCTION_BEGIN(aesECBEncrypt)
-LABEL(aesECBEncrypt)
-	subi r1,r1,32
-	stmw r24,0(r1)
-
-	mtctr r4
-
-	lis r12,_ae0@ha
-	la r12,_ae0@l(r12)
-
-LOCAL(02):
-	/* copy r3 into r4 */
-	mr r4,r3
-
-	eblock rk=r4 src=r6 label=LOCAL(03)
-
-	#if WORDS_BIGENDIAN
-	stw s0, 0(r5)
-	stw s1, 4(r5)
-	stw s2, 8(r5)
-	stw s3,12(r5)
-	#else
-	# error ppc little-endian mode not supported
-	#endif
-
-	addi r5,r5,16
-	addi r6,r6,16
-
-	bdnz LOCAL(02)
-
-	li r3,0
-	lmw r24,0(r1)
-	addi r1,r1,32
-	blr
-C_FUNCTION_END(aesECBEncrypt, LOCAL(aesECBEncrypt_size))
-
-
-C_FUNCTION_BEGIN(aesECBDecrypt)
-LABEL(aesECBDecrypt)
-	subi r1,r1,32
-	stmw r24,0(r1)
-
-	mtctr r4
-
-	lis r12,_ad0@ha
-	la r12,_ad0@l(r12)
-
-LOCAL(04):
-	/* copy r3 into r4 */
-	mr r4,r3
-
-	dblock rk=r4 src=r6 label=LOCAL(05)
-
-	#if WORDS_BIGENDIAN
-	stw s0, 0(r5)
-	stw s1, 4(r5)
-	stw s2, 8(r5)
-	stw s3,12(r5)
-	#else
-	# error ppc little-endian mode not supported
-	#endif
-
-	addi r5,r5,16
-	addi r6,r6,16
-
-	bdnz LOCAL(04)
-
-	li r3,0
-	lmw r24,0(r1)
-	addi r1,r1,32
-	blr
-C_FUNCTION_END(aesECBDecrypt, LOCAL(aesECBDecrypt_size))
diff --git a/beecrypt/gas/aesopt.ppc.m4 b/beecrypt/gas/aesopt.ppc.m4
new file mode 100644
index 000000000..f81f3a50a
--- /dev/null
+++ b/beecrypt/gas/aesopt.ppc.m4
@@ -0,0 +1,616 @@
+dnl  aesopt.ppc.m4
+dnl
+dnl  NOTE: Only works for big-endian PowerPC!
+dnl
+dnl  Copyright (c) 2003 Bob Deblier
+dnl 
+dnl  Author: Bob Deblier <bob.deblier@pandora.be>
+dnl 
+dnl  This library is free software; you can redistribute it and/or
+dnl  modify it under the terms of the GNU Lesser General Public
+dnl  License as published by the Free Software Foundation; either
+dnl  version 2.1 of the License, or (at your option) any later version.
+dnl 
+dnl  This library 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 GNU
+dnl  Lesser General Public License for more details.
+dnl 
+dnl  You should have received a copy of the GNU Lesser General Public
+dnl  License along with this library; if not, write to the Free Software
+dnl  Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
+
+include(config.m4)
+include(ASM_SRCDIR/ppc.m4)
+
+define(`s0',`r24')
+define(`s1',`r25')
+define(`s2',`r26')
+define(`s3',`r27')
+define(`t0',`r28')
+define(`t1',`r29')
+define(`t2',`r30')
+define(`t3',`r31')
+
+define(`sxrk',`
+ifelse(ASM_BIGENDIAN,yes,`
+	lwz s0, 0($2)
+	lwz s1, 4($2)
+	lwz s2, 8($2)
+	lwz s3,12($2)
+',`
+	li r0,0
+	lwbrx s0,$2,r0
+	li r0,4
+	lwbrx s1,$2,r0
+	li r0,8
+	lwbrx s2,$2,r0
+	li r0,13
+	lwbrx s0,$2,r0
+')
+	lwz r7, 0($1)
+	lwz r8, 4($1)
+	lwz r9, 8($1)
+	lwz r10,12($1)
+	xor s0,s0,r7
+	xor s1,s1,r8
+	xor s2,s2,r9
+	xor s3,s3,r10
+')
+
+define(`etfs',`
+	lwz t0,$2+ 0($1)
+	lwz t1,$2+ 4($1)
+	lwz t2,$2+ 8($1)
+	lwz t3,$2+12($1)
+
+	rlwinm r7,s0,10,22,29 
+	rlwinm r8,s1,10,22,29
+	rlwinm r9,s2,10,22,29
+	rlwinm r10,s3,10,22,29
+	lwzx r7,r7,r12
+	lwzx r8,r8,r12
+	lwzx r9,r9,r12
+	lwzx r10,r10,r12
+	xor t0,t0,r7
+	xor t1,t1,r8
+	xor t2,t2,r9
+	xor t3,t3,r10
+
+	la r12,1024(r12)
+
+	rlwinm r7,s1,18,22,29
+	rlwinm r8,s2,18,22,29
+	rlwinm r9,s3,18,22,29
+	rlwinm r10,s0,18,22,29
+	lwzx r7,r7,r12
+	lwzx r8,r8,r12
+	lwzx r9,r9,r12
+	lwzx r10,r10,r12
+	xor t0,t0,r7
+	xor t1,t1,r8
+	xor t2,t2,r9
+	xor t3,t3,r10
+
+	la r12,1024(r12)
+
+	rlwinm r7,s2,26,22,29
+	rlwinm r8,s3,26,22,29
+	rlwinm r9,s0,26,22,29
+	rlwinm r10,s1,26,22,29
+	lwzx r7,r7,r12
+	lwzx r8,r8,r12
+	lwzx r9,r9,r12
+	lwzx r10,r10,r12
+	xor t0,t0,r7
+	xor t1,t1,r8
+	xor t2,t2,r9
+	xor t3,t3,r10
+
+	la r12,1024(r12)
+
+	rlwinm r7,s3,2,22,29
+	rlwinm r8,s0,2,22,29
+	rlwinm r9,s1,2,22,29
+	rlwinm r10,s2,2,22,29
+	lwzx r7,r7,r12
+	lwzx r8,r8,r12
+	lwzx r9,r9,r12
+	lwzx r10,r10,r12
+	xor t0,t0,r7
+	xor t1,t1,r8
+	xor t2,t2,r9
+	xor t3,t3,r10
+
+	la r12,-3072(r12)
+')
+
+define(`esft',`
+	lwz s0,$2+ 0($1)
+	lwz s1,$2+ 4($1)
+	lwz s2,$2+ 8($1)
+	lwz s3,$2+12($1)
+
+	rlwinm r7,t0,10,22,29
+	rlwinm r8,t1,10,22,29
+	rlwinm r9,t2,10,22,29
+	rlwinm r10,t3,10,22,29
+	lwzx r7,r7,r12
+	lwzx r8,r8,r12
+	lwzx r9,r9,r12
+	lwzx r10,r10,r12
+	xor s0,s0,r7
+	xor s1,s1,r8
+	xor s2,s2,r9
+	xor s3,s3,r10
+
+	la r12,1024(r12)
+
+	rlwinm r7,t1,18,22,29
+	rlwinm r8,t2,18,22,29
+	rlwinm r9,t3,18,22,29
+	rlwinm r10,t0,18,22,29
+	lwzx r7,r7,r12
+	lwzx r8,r8,r12
+	lwzx r9,r9,r12
+	lwzx r10,r10,r12
+	xor s0,s0,r7
+	xor s1,s1,r8
+	xor s2,s2,r9
+	xor s3,s3,r10
+
+	la r12,1024(r12)
+
+	rlwinm r7,t2,26,22,29
+	rlwinm r8,t3,26,22,29
+	rlwinm r9,t0,26,22,29
+	rlwinm r10,t1,26,22,29
+	lwzx r7,r7,r12
+	lwzx r8,r8,r12
+	lwzx r9,r9,r12
+	lwzx r10,r10,r12
+	xor s0,s0,r7
+	xor s1,s1,r8
+	xor s2,s2,r9
+	xor s3,s3,r10
+
+	la r12,1024(r12)
+
+	rlwinm r7,t3,2,22,29
+	rlwinm r8,t0,2,22,29
+	rlwinm r9,t1,2,22,29
+	rlwinm r10,t2,2,22,29
+	lwzx r7,r7,r12
+	lwzx r8,r8,r12
+	lwzx r9,r9,r12
+	lwzx r10,r10,r12
+	xor s0,s0,r7
+	xor s1,s1,r8
+	xor s2,s2,r9
+	xor s3,s3,r10
+
+	la r12,-3072(r12)
+')
+
+define(`elr',`
+	lwz s0, 0($1)
+	lwz s1, 4($1)
+	lwz s2, 8($1)
+	lwz s3,12($1)
+
+	la r12,4096(r12)
+
+	rlwinm r7,t0,10,22,29
+	rlwinm r8,t1,10,22,29
+	rlwinm r9,t2,10,22,29
+	rlwinm r10,t3,10,22,29
+	lwzx r7,r7,r12
+	lwzx r8,r8,r12
+	lwzx r9,r9,r12
+	lwzx r10,r10,r12
+	rlwinm r7,r7,0,0,7
+	rlwinm r8,r8,0,0,7
+	rlwinm r9,r9,0,0,7
+	rlwinm r10,r10,0,0,7
+	xor s0,s0,r7
+	xor s1,s1,r8
+	xor s2,s2,r9
+	xor s3,s3,r10
+
+	rlwinm r7,t1,18,22,29
+	rlwinm r8,t2,18,22,29
+	rlwinm r9,t3,18,22,29
+	rlwinm r10,t0,18,22,29
+	lwzx r7,r7,r12
+	lwzx r8,r8,r12
+	lwzx r9,r9,r12
+	lwzx r10,r10,r12
+	rlwinm r7,r7,0,8,15
+	rlwinm r8,r8,0,8,15
+	rlwinm r9,r9,0,8,15
+	rlwinm r10,r10,0,8,15
+	xor s0,s0,r7
+	xor s1,s1,r8
+	xor s2,s2,r9
+	xor s3,s3,r10
+
+	rlwinm r7,t2,26,22,29
+	rlwinm r8,t3,26,22,29
+	rlwinm r9,t0,26,22,29
+	rlwinm r10,t1,26,22,29
+	lwzx r7,r7,r12
+	lwzx r8,r8,r12
+	lwzx r9,r9,r12
+	lwzx r10,r10,r12
+	rlwinm r7,r7,0,16,23
+	rlwinm r8,r8,0,16,23
+	rlwinm r9,r9,0,16,23
+	rlwinm r10,r10,0,16,23
+	xor s0,s0,r7
+	xor s1,s1,r8
+	xor s2,s2,r9
+	xor s3,s3,r10
+
+	rlwinm r7,t3,2,22,29
+	rlwinm r8,t0,2,22,29
+	rlwinm r9,t1,2,22,29
+	rlwinm r10,t2,2,22,29
+	lwzx r7,r7,r12
+	lwzx r8,r8,r12
+	lwzx r9,r9,r12
+	lwzx r10,r10,r12
+	rlwinm r7,r7,0,24,31
+	rlwinm r8,r8,0,24,31
+	rlwinm r9,r9,0,24,31
+	rlwinm r10,r10,0,24,31
+	xor s0,s0,r7
+	xor s1,s1,r8
+	xor s2,s2,r9
+	xor s3,s3,r10
+
+	la r12,-4096(r12)
+')
+
+define(`eblock',`
+	sxrk($1,$2)
+
+	etfs($1,16)
+	esft($1,32)
+	etfs($1,48)
+	esft($1,64)
+	etfs($1,80)
+	esft($1,96)
+	etfs($1,112)
+	esft($1,128)
+	etfs($1,144)
+
+	lwz r11,256($1)
+	cmpwi r11,10
+	beq $3
+
+	esft($1,160)
+	etfs($1,176)
+
+	cmpwi r11,12
+	beq $3
+
+	esft($1,192)
+	etfs($1,208)
+
+$3:
+	slwi r11,r11,4
+	add $1,$1,r11
+
+	elr($1)
+')
+
+define(`dtfs',`
+	lwz t0,$2+ 0($1)
+	lwz t1,$2+ 4($1)
+	lwz t2,$2+ 8($1)
+	lwz t3,$2+12($1)
+
+	rlwinm r7,s0,10,22,29
+	rlwinm r8,s1,10,22,29
+	rlwinm r9,s2,10,22,29
+	rlwinm r10,s3,10,22,29
+	lwzx r7,r7,r12
+	lwzx r8,r8,r12
+	lwzx r9,r9,r12
+	lwzx r10,r10,r12
+	xor t0,t0,r7
+	xor t1,t1,r8
+	xor t2,t2,r9
+	xor t3,t3,r10
+
+	la r12,1024(r12)
+
+	rlwinm r7,s3,18,22,29
+	rlwinm r8,s0,18,22,29
+	rlwinm r9,s1,18,22,29
+	rlwinm r10,s2,18,22,29
+
+	lwzx r7,r7,r12
+	lwzx r8,r8,r12
+	lwzx r9,r9,r12
+	lwzx r10,r10,r12
+	xor t0,t0,r7
+	xor t1,t1,r8
+	xor t2,t2,r9
+	xor t3,t3,r10
+
+	la r12,1024(r12)
+
+	rlwinm r7,s2,26,22,29
+	rlwinm r8,s3,26,22,29
+	rlwinm r9,s0,26,22,29
+	rlwinm r10,s1,26,22,29
+	lwzx r7,r7,r12
+	lwzx r8,r8,r12
+	lwzx r9,r9,r12
+	lwzx r10,r10,r12
+	xor t0,t0,r7
+	xor t1,t1,r8
+	xor t2,t2,r9
+	xor t3,t3,r10
+
+	la r12,1024(r12)
+
+	rlwinm r7,s1,2,22,29
+	rlwinm r8,s2,2,22,29
+	rlwinm r9,s3,2,22,29
+	rlwinm r10,s0,2,22,29
+	lwzx r7,r7,r12
+	lwzx r8,r8,r12
+	lwzx r9,r9,r12
+	lwzx r10,r10,r12
+	xor t0,t0,r7
+	xor t1,t1,r8
+	xor t2,t2,r9
+	xor t3,t3,r10
+
+	la r12,-3072(r12)
+')
+
+define(`dsft',`
+	lwz s0,$2+ 0($1)
+	lwz s1,$2+ 4($1)
+	lwz s2,$2+ 8($1)
+	lwz s3,$2+12($1)
+
+	rlwinm r7,t0,10,22,29
+	rlwinm r8,t1,10,22,29
+	rlwinm r9,t2,10,22,29
+	rlwinm r10,t3,10,22,29
+	lwzx r7,r7,r12
+	lwzx r8,r8,r12
+	lwzx r9,r9,r12
+	lwzx r10,r10,r12
+	xor s0,s0,r7
+	xor s1,s1,r8
+	xor s2,s2,r9
+	xor s3,s3,r10
+
+	la r12,1024(r12)
+
+	rlwinm r7,t3,18,22,29
+	rlwinm r8,t0,18,22,29
+	rlwinm r9,t1,18,22,29
+	rlwinm r10,t2,18,22,29
+	lwzx r7,r7,r12
+	lwzx r8,r8,r12
+	lwzx r9,r9,r12
+	lwzx r10,r10,r12
+	xor s0,s0,r7
+	xor s1,s1,r8
+	xor s2,s2,r9
+	xor s3,s3,r10
+
+	la r12,1024(r12)
+
+	rlwinm r7,t2,26,22,29
+	rlwinm r8,t3,26,22,29
+	rlwinm r9,t0,26,22,29
+	rlwinm r10,t1,26,22,29
+	lwzx r7,r7,r12
+	lwzx r8,r8,r12
+	lwzx r9,r9,r12
+	lwzx r10,r10,r12
+	xor s0,s0,r7
+	xor s1,s1,r8
+	xor s2,s2,r9
+	xor s3,s3,r10
+
+	la r12,1024(r12)
+
+	rlwinm r7,t1,2,22,29
+	rlwinm r8,t2,2,22,29
+	rlwinm r9,t3,2,22,29
+	rlwinm r10,t0,2,22,29
+	lwzx r7,r7,r12
+	lwzx r8,r8,r12
+	lwzx r9,r9,r12
+	lwzx r10,r10,r12
+	xor s0,s0,r7
+	xor s1,s1,r8
+	xor s2,s2,r9
+	xor s3,s3,r10
+
+	la r12,-3072(r12)
+')
+
+define(`dlr',`
+	lwz s0, 0($1)
+	lwz s1, 4($1)
+	lwz s2, 8($1)
+	lwz s3,12($1)
+
+	la r12,4096(r12)
+
+	rlwinm r7,t0,10,22,29
+	rlwinm r8,t1,10,22,29
+	rlwinm r9,t2,10,22,29
+	rlwinm r10,t3,10,22,29
+	lwzx r7,r7,r12
+	lwzx r8,r8,r12
+	lwzx r9,r9,r12
+	lwzx r10,r10,r12
+	rlwinm r7,r7,0,0,7
+	rlwinm r8,r8,0,0,7
+	rlwinm r9,r9,0,0,7
+	rlwinm r10,r10,0,0,7
+	xor s0,s0,r7
+	xor s1,s1,r8
+	xor s2,s2,r9
+	xor s3,s3,r10
+
+	rlwinm r7,t3,18,22,29
+	rlwinm r8,t0,18,22,29
+	rlwinm r9,t1,18,22,29
+	rlwinm r10,t2,18,22,29
+	lwzx r7,r7,r12
+	lwzx r8,r8,r12
+	lwzx r9,r9,r12
+	lwzx r10,r10,r12
+	rlwinm r7,r7,0,8,15
+	rlwinm r8,r8,0,8,15
+	rlwinm r9,r9,0,8,15
+	rlwinm r10,r10,0,8,15
+	xor s0,s0,r7
+	xor s1,s1,r8
+	xor s2,s2,r9
+	xor s3,s3,r10
+
+	rlwinm r7,t2,26,22,29
+	rlwinm r8,t3,26,22,29
+	rlwinm r9,t0,26,22,29
+	rlwinm r10,t1,26,22,29
+	lwzx r7,r7,r12
+	lwzx r8,r8,r12
+	lwzx r9,r9,r12
+	lwzx r10,r10,r12
+	rlwinm r7,r7,0,16,23
+	rlwinm r8,r8,0,16,23
+	rlwinm r9,r9,0,16,23
+	rlwinm r10,r10,0,16,23
+	xor s0,s0,r7
+	xor s1,s1,r8
+	xor s2,s2,r9
+	xor s3,s3,r10
+
+	rlwinm r7,t1,2,22,29
+	rlwinm r8,t2,2,22,29
+	rlwinm r9,t3,2,22,29
+	rlwinm r10,t0,2,22,29
+	lwzx r7,r7,r12
+	lwzx r8,r8,r12
+	lwzx r9,r9,r12
+	lwzx r10,r10,r12
+	rlwinm r7,r7,0,24,31
+	rlwinm r8,r8,0,24,31
+	rlwinm r9,r9,0,24,31
+	rlwinm r10,r10,0,24,31
+	xor s0,s0,r7
+	xor s1,s1,r8
+	xor s2,s2,r9
+	xor s3,s3,r10
+
+	la r12,-4096(r12)
+')
+
+define(`dblock',`
+	sxrk($1,$2)
+
+	dtfs($1,16)
+	dsft($1,32)
+	dtfs($1,48)
+	dsft($1,64)
+	dtfs($1,80)
+	dsft($1,96)
+	dtfs($1,112)
+	dsft($1,128)
+	dtfs($1,144)
+
+	lwz r11,256($1)
+	cmpwi r11,10
+	beq $3
+
+	dsft($1,160)
+	dtfs($1,176)
+
+	cmpwi r11,12
+	beq $3
+
+	dsft($1,192)
+	dtfs($1,208)
+
+$3:
+	slwi r11,r11,4
+	add $1,$1,r11
+
+	dlr($1)
+')
+
+EXTERNAL_VARIABLE(_ae0)
+EXTERNAL_VARIABLE(_ad0)
+
+C_FUNCTION_BEGIN(aesEncrypt)
+	subi r1,r1,32
+	stmw r24,0(r1)
+
+	LOAD_ADDRESS(_ae0,r12)
+
+	eblock(r3,r5,LOCAL(00))
+
+ifelse(ASM_BIGENDIAN,yes,`
+	stw s0, 0(r4)
+	stw s1, 4(r4)
+	stw s2, 8(r4)
+	stw s3,12(r4)
+',`
+	li r0,0
+	stwbrx s0,r4,r0
+	li r0,4
+	stwbrx s1,r4,r0
+	li r0,8
+	stwbrx s2,r4,r0
+	li r0,12
+	stwbrx s3,r4,r0
+')
+
+	li r3,0
+	lmw r24,0(r1)
+	addi r1,r1,32
+	blr
+C_FUNCTION_END(aesEncrypt)
+
+
+C_FUNCTION_BEGIN(aesDecrypt)
+	subi r1,r1,32
+	stmw r24,0(r1)
+
+	LOAD_ADDRESS(_ad0,r12)
+
+	dblock(r3,r5,LOCAL(01))
+
+ifelse(ASM_BIGENDIAN,yes,`
+	stw s0, 0(r4)
+	stw s1, 4(r4)
+	stw s2, 8(r4)
+	stw s3,12(r4)
+',`
+	li r0,0
+	stwbrx s0,r4,r0
+	li r0,4
+	stwbrx s1,r4,r0
+	li r0,8
+	stwbrx s2,r4,r0
+	li r0,12
+	stwbrx s3,r4,r0
+')
+
+	li r3,0
+	lmw r24,0(r1)
+	addi r1,r1,32
+	blr
+C_FUNCTION_END(aesDecrypt)
diff --git a/beecrypt/gas/alpha.m4 b/beecrypt/gas/alpha.m4
new file mode 100644
index 000000000..49366dae0
--- /dev/null
+++ b/beecrypt/gas/alpha.m4
@@ -0,0 +1,34 @@
+dnl  alpha.m4
+dnl
+dnl  Copyright (c) 2003 Bob Deblier
+dnl 
+dnl  Author: Bob Deblier <bob.deblier@pandora.be>
+dnl 
+dnl  This library is free software; you can redistribute it and/or
+dnl  modify it under the terms of the GNU Lesser General Public
+dnl  License as published by the Free Software Foundation; either
+dnl  version 2.1 of the License, or (at your option) any later version.
+dnl 
+dnl  This library 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 GNU
+dnl  Lesser General Public License for more details.
+dnl 
+dnl  You should have received a copy of the GNU Lesser General Public
+dnl  License along with this library; if not, write to the Free Software
+dnl  Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
+
+undefine(`C_FUNCTION_BEGIN')
+define(C_FUNCTION_BEGIN,`
+	.text
+	.align 5
+	.globl $1
+	.ent $1
+$1:
+	.frame `$'sp, 0, `$'26
+	.prologue 0
+')
+undefine(`C_FUNCTION_END')
+define(C_FUNCTION_END,`
+	.end $1
+')
diff --git a/beecrypt/gas/asmdefs.m4 b/beecrypt/gas/asmdefs.m4
new file mode 100644
index 000000000..f8341c95f
--- /dev/null
+++ b/beecrypt/gas/asmdefs.m4
@@ -0,0 +1,41 @@
+dnl  asmdefs.m4
+dnl
+dnl  Copyright (c) 2003 Bob Deblier
+dnl 
+dnl  Author: Bob Deblier <bob.deblier@pandora.be>
+dnl 
+dnl  This library is free software; you can redistribute it and/or
+dnl  modify it under the terms of the GNU Lesser General Public
+dnl  License as published by the Free Software Foundation; either
+dnl  version 2.1 of the License, or (at your option) any later version.
+dnl 
+dnl  This library 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 GNU
+dnl  Lesser General Public License for more details.
+dnl 
+dnl  You should have received a copy of the GNU Lesser General Public
+dnl  License along with this library; if not, write to the Free Software
+dnl  Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
+
+ifelse(substr(ASM_OS,0,5),linux,`
+define(USE_SIZE_DIRECTIVE,yes)
+')
+
+define(SYMNAME,`GSYM_PREFIX`$1'')
+define(LOCAL,`LSYM_PREFIX`$1'')
+
+define(C_FUNCTION_BEGIN,`
+	TEXTSEG
+	GLOBL SYMNAME($1)
+SYMNAME($1):
+')
+
+ifelse(USE_SIZE_DIRECTIVE,yes,`
+define(C_FUNCTION_END,`
+LOCAL($1)_size:
+	.size SYMNAME($1), LOCAL($1)_size  - SYMNAME($1)
+')
+',`
+define(C_FUNCTION_END,`')
+')
diff --git a/beecrypt/gas/blowfishopt.i586.S b/beecrypt/gas/blowfishopt.i586.S
deleted file mode 100644
index 69b89bb72..000000000
--- a/beecrypt/gas/blowfishopt.i586.S
+++ /dev/null
@@ -1,178 +0,0 @@
-/*
- * blowfishopt.i586.S
- *
- * Assembler optimized blowfish routines for Intel Pentium processors
- *
- * Compile target is GNU Assembler
- *
- * Copyright (c) 2000, 2001 Virtual Unlimited B.V.
- *
- * Author: Bob Deblier <bob@virtualunlimited.com>
- *
- * This library is free software; you can redistribute it and/or
- * modify it under the terms of the GNU Lesser General Public
- * License as published by the Free Software Foundation; either
- * version 2.1 of the License, or (at your option) any later version.
- *
- * This library 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
- * Lesser General Public License for more details.
- *
- * You should have received a copy of the GNU Lesser General Public
- * License along with this library; if not, write to the Free Software
- * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
- *
- */
-
-#include "beecrypt.gas.h"
-
-	.file "blowfishopt.i586.S"
-
-	.text
-
-	.macro  etworounds	p # bp in %esi, xl and xr in %ecx and %edx, %eax and %ebx clear
-	xorl 0+\p(%esi),%ecx
-	roll $16,%ecx
-	movb %ch,%al
-	movb %cl,%bl
-	roll $16,%ecx
-	movl 0x000+72(%esi,%eax,4),%edi
-	addl 0x400+72(%esi,%ebx,4),%edi
-	movb %ch,%al
-	movb %cl,%bl
-	xorl 0x800+72(%esi,%eax,4),%edi
-	addl 0xC00+72(%esi,%ebx,4),%edi
-	xorl %edi,%edx
-	xorl 4+\p(%esi),%edx
-	roll $16,%edx
-	movb %dh,%al
-	movb %dl,%bl
-	roll $16,%edx
-	movl 0x000+72(%esi,%eax,4),%edi
-	addl 0x400+72(%esi,%ebx,4),%edi
-	movb %dh,%al
-	movb %dl,%bl
-	xorl 0x800+72(%esi,%eax,4),%edi
-	addl 0xC00+72(%esi,%ebx,4),%edi
-	xorl %edi,%ecx
-	.endm
-
-	.macro  dtworounds	p # bp in %esi, xl and xr in %ecx and %edx, %eax and %ebx clear
-	xorl 4+\p(%esi),%ecx
-	roll $16,%ecx
-	movb %ch,%al
-	movb %cl,%bl
-	roll $16,%ecx
-	movl 0x000+72(%esi,%eax,4),%edi
-	addl 0x400+72(%esi,%ebx,4),%edi
-	movb %ch,%al
-	movb %cl,%bl
-	xorl 0x800+72(%esi,%eax,4),%edi
-	addl 0xC00+72(%esi,%ebx,4),%edi
-	xorl %edi,%edx
-	xorl 0+\p(%esi),%edx
-	roll $16,%edx
-	movb %dh,%al
-	movb %dl,%bl
-	roll $16,%edx
-	movl 0x000+72(%esi,%eax,4),%edi
-	addl 0x400+72(%esi,%ebx,4),%edi
-	movb %dh,%al
-	movb %dl,%bl
-	xorl 0x800+72(%esi,%eax,4),%edi
-	addl 0xC00+72(%esi,%ebx,4),%edi
-	xorl %edi,%ecx
-	.endm
-
-C_FUNCTION_BEGIN(blowfishEncrypt)
-LABEL(blowfishEncrypt)
-	/* parameter one is the blowfish parameters; need to extract bp and set it up in esi */
-	pushl %edi
-	pushl %esi
-	pushl %ebx
-
-	movl 16(%esp),%esi # esi now contains bp
-	movl 24(%esp),%edi # edi now contains src
-
-	xorl %eax,%eax
-	xorl %ebx,%ebx
-
-	movl 0(%edi),%ecx
-	movl 4(%edi),%edx
-
-	bswap %ecx
-	bswap %edx
-
-	etworounds p= 0
-	etworounds p= 8
-	etworounds p=16
-	etworounds p=24
-	etworounds p=32
-	etworounds p=40
-	etworounds p=48
-	etworounds p=56
-
-	movl 20(%esp),%edi	# edi now contains dst
-	xorl 64(%esi),%ecx
-	xorl 68(%esi),%edx
-
-	bswap %ecx
-	bswap %edx
-
-	movl %ecx,4(%edi)
-	movl %edx,0(%edi)
-
-	xorl %eax,%eax
-	popl %ebx
-	popl %esi
-	popl %edi
-	ret
-C_FUNCTION_END(blowfishEncrypt, LOCAL(blowfishEncrypt_size))
-
-
-C_FUNCTION_BEGIN(blowfishDecrypt)
-LABEL(blowfishDecrypt)
-	/* parameter one is the blowfish parameters; need to extract bp and set it up in ebp */
-	pushl %edi
-	pushl %esi
-	pushl %ebx
-
-	movl 16(%esp),%esi # esi now contains bp
-	movl 24(%esp),%edi # edi now contains dst
-
-	xorl %eax,%eax
-	xorl %ebx,%ebx
-
-	movl 0(%edi),%ecx
-	movl 4(%edi),%edx
-
-	bswap %ecx
-	bswap %edx
-	
-	dtworounds p=64
-	dtworounds p=56
-	dtworounds p=48
-	dtworounds p=40
-	dtworounds p=32
-	dtworounds p=24
-	dtworounds p=16
-	dtworounds p= 8
-
-	movl 20(%esp),%edi	# edi now contains dst
-	xorl 4(%esi),%ecx
-	xorl 0(%esi),%edx
-	
-	bswap %ecx
-	bswap %edx
-	
-	movl %ecx,4(%edi)
-	movl %edx,0(%edi)
-
-	xorl %eax,%eax
-
-	popl %ebx
-	popl %esi
-	popl %edi
-	ret
-C_FUNCTION_END(blowfishDecrypt, LOCAL(blowfishDecrypt_size))
diff --git a/beecrypt/gas/blowfishopt.i586.m4 b/beecrypt/gas/blowfishopt.i586.m4
new file mode 100644
index 000000000..4233738fc
--- /dev/null
+++ b/beecrypt/gas/blowfishopt.i586.m4
@@ -0,0 +1,162 @@
+dnl  blowfishopt.i586.m4
+dnl
+dnl  Copyright (c) 2003 Bob Deblier
+dnl 
+dnl  Author: Bob Deblier <bob.deblier@pandora.be>
+dnl 
+dnl  This library is free software; you can redistribute it and/or
+dnl  modify it under the terms of the GNU Lesser General Public
+dnl  License as published by the Free Software Foundation; either
+dnl  version 2.1 of the License, or (at your option) any later version.
+dnl 
+dnl  This library 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 GNU
+dnl  Lesser General Public License for more details.
+dnl 
+dnl  You should have received a copy of the GNU Lesser General Public
+dnl  License along with this library; if not, write to the Free Software
+dnl  Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
+
+include(config.m4)
+include(ASM_SRCDIR/x86.m4)
+
+dnl during this macro we assume:
+dnl	bp in %esi, xl and xr in %ecx and %edx, %eax and %ebx clear
+
+define(`etworounds',`
+	xorl $1+0(%esi),%ecx
+	roll `$'16,%ecx
+	movzx %ch,%eax
+	movzx %cl,%ebx
+	roll `$'16,%ecx
+	movl 0x000+72(%esi,%eax,4),%edi
+	addl 0x400+72(%esi,%ebx,4),%edi
+	movzx %ch,%eax
+	movzx %cl,%ebx
+	xorl 0x800+72(%esi,%eax,4),%edi
+	addl 0xC00+72(%esi,%ebx,4),%edi
+	xorl %edi,%edx
+	xorl $1+4(%esi),%edx
+	roll `$'16,%edx
+	movzx %dh,%eax
+	movzx %dl,%ebx
+	roll `$'16,%edx
+	movl 0x000+72(%esi,%eax,4),%edi
+	addl 0x400+72(%esi,%ebx,4),%edi
+	movzx %dh,%eax
+	movzx %dl,%ebx
+	xorl 0x800+72(%esi,%eax,4),%edi
+	addl 0xC00+72(%esi,%ebx,4),%edi
+	xorl %edi,%ecx
+')
+
+dnl bp in %esi, xl and xr in %ecx and %edx, %eax and %ebx clear
+define(`dtworounds',`
+	xorl $1+4(%esi),%ecx
+	roll `$'16,%ecx
+	movzx %ch,%eax
+	movzx %cl,%ebx
+	roll `$'16,%ecx
+	movl 0x000+72(%esi,%eax,4),%edi
+	addl 0x400+72(%esi,%ebx,4),%edi
+	movzx %ch,%eax
+	movzx %cl,%ebx
+	xorl 0x800+72(%esi,%eax,4),%edi
+	addl 0xC00+72(%esi,%ebx,4),%edi
+	xorl %edi,%edx
+	xorl $1+0(%esi),%edx
+	roll `$'16,%edx
+	movzx %dh,%eax
+	movzx %dl,%ebx
+	roll `$'16,%edx
+	movl 0x000+72(%esi,%eax,4),%edi
+	addl 0x400+72(%esi,%ebx,4),%edi
+	movzx %dh,%eax
+	movzx %dl,%ebx
+	xorl 0x800+72(%esi,%eax,4),%edi
+	addl 0xC00+72(%esi,%ebx,4),%edi
+	xorl %edi,%ecx
+')
+
+C_FUNCTION_BEGIN(blowfishEncrypt)
+	pushl %edi
+	pushl %esi
+	pushl %ebx
+
+	movl 16(%esp),%esi
+	movl 24(%esp),%edi
+
+	movl 0(%edi),%ecx
+	movl 4(%edi),%edx
+
+	bswap %ecx
+	bswap %edx
+
+	etworounds(0)
+	etworounds(8)
+	etworounds(16)
+	etworounds(24)
+	etworounds(32)
+	etworounds(40)
+	etworounds(48)
+	etworounds(56)
+
+	movl 20(%esp),%edi
+	xorl 64(%esi),%ecx
+	xorl 68(%esi),%edx
+
+	bswap %ecx
+	bswap %edx
+
+	movl %ecx,4(%edi)
+	movl %edx,0(%edi)
+
+	xorl %eax,%eax
+	popl %ebx
+	popl %esi
+	popl %edi
+	ret
+C_FUNCTION_END(blowfishEncrypt)
+
+
+C_FUNCTION_BEGIN(blowfishDecrypt)
+	pushl %edi
+	pushl %esi
+	pushl %ebx
+
+	movl 16(%esp),%esi
+	movl 24(%esp),%edi
+
+	movl 0(%edi),%ecx
+	movl 4(%edi),%edx
+
+	bswap %ecx
+	bswap %edx
+	
+	dtworounds(64)
+	dtworounds(56)
+	dtworounds(48)
+	dtworounds(40)
+	dtworounds(32)
+	dtworounds(24)
+	dtworounds(16)
+	dtworounds(8)
+
+	movl 20(%esp),%edi
+	xorl 4(%esi),%ecx
+	xorl 0(%esi),%edx
+	
+	bswap %ecx
+	bswap %edx
+	
+	movl %ecx,4(%edi)
+	movl %edx,0(%edi)
+
+	xorl %eax,%eax
+
+	popl %ebx
+	popl %esi
+	popl %edi
+	ret
+C_FUNCTION_END(blowfishDecrypt)
diff --git a/beecrypt/gas/blowfishopt.powerpc.S b/beecrypt/gas/blowfishopt.powerpc.S
deleted file mode 100644
index 28076f061..000000000
--- a/beecrypt/gas/blowfishopt.powerpc.S
+++ /dev/null
@@ -1,242 +0,0 @@
-/*
- * blowfishopt.powerpc.asm
- *
- * Assembler optimized Blowfish routines for PowerPC processors
- *
- * Compile target is GNU Assembler
- *
- * Copyright (c) 2002 Bob Deblier <bob@virtualunlimited.com>
- *
- * This library is free software; you can redistribute it and/or
- * modify it under the terms of the GNU Lesser General Public
- * License as published by the Free Software Foundation; either
- * version 2.1 of the License, or (at your option) any later version.
- *
- * This library 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
- * Lesser General Public License for more details.
- *
- * You should have received a copy of the GNU Lesser General Public
- * License along with this library; if not, write to the Free Software
- * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
- *
- */
-
-#include "beecrypt.gas.h"
-
-	.file "blowfish.powerpc.S"
-
-	.text
-
-	.macro round	xl xr offset
-	lwz r9,\offset(r3)
-	xor \xl,\xl,r9
-	rlwinm r9,\xl,10,22,29
-	rlwinm r10,\xl,18,22,29
-	lwzx r9,r9,r28
-	lwzx r10,r10,r29
-	rlwinm r11,\xl,26,22,29
-	add r9,r9,r10
-	lwzx r11,r11,r30
-	rlwinm r12,\xl,2,22,29
-	xor r9,r9,r11
-	lwzx r12,r12,r31
-	add r9,r9,r12
-	xor \xr,\xr,r9
-	.endm
-
-	.macro eblock
-	round xl=r7 xr=r8 offset=0
-	round xl=r8 xr=r7 offset=4
-	round xl=r7 xr=r8 offset=8
-	round xl=r8 xr=r7 offset=12
-	round xl=r7 xr=r8 offset=16
-	round xl=r8 xr=r7 offset=20
-	round xl=r7 xr=r8 offset=24
-	round xl=r8 xr=r7 offset=28
-	round xl=r7 xr=r8 offset=32
-	round xl=r8 xr=r7 offset=36
-	round xl=r7 xr=r8 offset=40
-	round xl=r8 xr=r7 offset=44
-	round xl=r7 xr=r8 offset=48
-	round xl=r8 xr=r7 offset=52
-	round xl=r7 xr=r8 offset=56
-	round xl=r8 xr=r7 offset=60
-
-	lwz r9,64(r3)
-	lwz r10,68(r3)
-	xor r7,r7,r9
-	xor r8,r8,r10
-	.endm
-
-	.macro dblock
-	round xl=r7 xr=r8 offset=68
-	round xl=r8 xr=r7 offset=64
-	round xl=r7 xr=r8 offset=60
-	round xl=r8 xr=r7 offset=56
-	round xl=r7 xr=r8 offset=52
-	round xl=r8 xr=r7 offset=48
-	round xl=r7 xr=r8 offset=44
-	round xl=r8 xr=r7 offset=40
-	round xl=r7 xr=r8 offset=36
-	round xl=r8 xr=r7 offset=32
-	round xl=r7 xr=r8 offset=28
-	round xl=r8 xr=r7 offset=24
-	round xl=r7 xr=r8 offset=20
-	round xl=r8 xr=r7 offset=16
-	round xl=r7 xr=r8 offset=12
-	round xl=r8 xr=r7 offset=8
-
-	lwz r9,4(r3)
-	lwz r10,0(r3)
-	xor r7,r7,r9
-	xor r8,r8,r10
-	.endm
-
-C_FUNCTION_BEGIN(blowfishEncrypt)
-LABEL(blowfishEncrypt)
-	la r1,-16(r1)
-	stmw r28,0(r1)
-
-	la r28,72(r3)
-	la r29,1096(r3)
-	la r30,2120(r3)
-	la r31,3144(r3)
-
-	#if WORDS_BIGENDIAN
-	lwz r7,0(r5)
-	lwz r8,4(r5)
-	#else
-	# error ppc little-endian mode not supported
-	#endif
-	
-	eblock
-
-	#if WORDS_BIGENDIAN
-	stw r7,4(r4)
-	stw r8,0(r4)
-	#else
-	# error ppc little-endian mode not supported
-	#endif
-
-	li r3,0
-	lmw r28,0(r1)
-	la r1,16(r1)
-	blr
-C_FUNCTION_END(blowfishEncrypt, LOCAL(blowfishEncrypt_size))
-
-
-C_FUNCTION_BEGIN(blowfishDecrypt)
-LABEL(blowfishDecrypt)
-	la r1,-16(r1)
-	stmw r28,0(r1)
-
-	la r28,72(r3)
-	la r29,1096(r3)
-	la r30,2120(r3)
-	la r31,3144(r3)
-
-	#if WORDS_BIGENDIAN
-	lwz r7,0(r5)
-	lwz r8,4(r5)
-	#else
-	# error ppc little-endian mode not supported
-	#endif
-	
-	dblock
-
-	#if WORDS_BIGENDIAN
-	stw r7,4(r4)
-	stw r8,0(r4)
-	#else
-	# error ppc little-endian mode not supported
-	#endif
-
-	li r3,0
-	lmw r28,0(r1)
-	la r1,16(r1)
-	blr
-C_FUNCTION_END(blowfishDecrypt, LOCAL(blowfishDecrypt_size))
-
-
-C_FUNCTION_BEGIN(blowfishECBEncrypt)
-LABEL(blowfishECBEncrypt)
-	la r1,-16(r1)
-	stmw r28,0(r1)
-
-	mtctr r4
-
-	la r28,72(r3)
-	la r29,1096(r3)
-	la r30,2120(r3)
-	la r31,3144(r3)
-
-LOCAL(00):
-	#if WORDS_BIGENDIAN
-	lwz r7,0(r6)
-	lwz r8,4(r6)
-	#else
-	# error ppc little-endian mode not supported
-	#endif
-	
-	eblock
-
-	#if WORDS_BIGENDIAN
-	stw r7,4(r5)
-	stw r8,0(r5)
-	#else
-	# error ppc little-endian mode not supported
-	#endif
-
-	la r5,8(r5)
-	la r6,8(r6)
-
-	bdnz LOCAL(00)
-
-	li r3,0
-	lmw r28,0(r1)
-	la r1,16(r1)
-	blr
-C_FUNCTION_END(blowfishECBEncrypt, LOCAL(blowfishECBEncrypt_size))
-
-
-C_FUNCTION_BEGIN(blowfishECBDecrypt)
-LABEL(blowfishECBDecrypt)
-	la r1,-16(r1)
-	stmw r28,0(r1)
-
-	mtctr r4
-
-	la r28,72(r3)
-	la r29,1096(r3)
-	la r30,2120(r3)
-	la r31,3144(r3)
-
-LOCAL(01):
-	#if WORDS_BIGENDIAN
-	lwz r7,0(r6)
-	lwz r8,4(r6)
-	#else
-	# error ppc little-endian mode not supported
-	#endif
-	
-	dblock
-
-	#if WORDS_BIGENDIAN
-	stw r7,4(r5)
-	stw r8,0(r5)
-	#else
-	# error ppc little-endian mode not supported
-	#endif
-
-	la r5,8(r5)
-	la r6,8(r6)
-
-	bdnz LOCAL(01)
-
-	li r3,0
-	lmw r28,0(r1)
-	la r1,16(r1)
-	blr
-C_FUNCTION_END(blowfishECBDecrypt, LOCAL(blowfishECBDecrypt_size))
diff --git a/beecrypt/gas/blowfishopt.ppc.m4 b/beecrypt/gas/blowfishopt.ppc.m4
new file mode 100644
index 000000000..74214aa9c
--- /dev/null
+++ b/beecrypt/gas/blowfishopt.ppc.m4
@@ -0,0 +1,161 @@
+dnl  blowfishopt.ppc.m4
+dnl
+dnl  Copyright (c) 2003 Bob Deblier
+dnl 
+dnl  Author: Bob Deblier <bob.deblier@pandora.be>
+dnl 
+dnl  This library is free software; you can redistribute it and/or
+dnl  modify it under the terms of the GNU Lesser General Public
+dnl  License as published by the Free Software Foundation; either
+dnl  version 2.1 of the License, or (at your option) any later version.
+dnl 
+dnl  This library 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 GNU
+dnl  Lesser General Public License for more details.
+dnl 
+dnl  You should have received a copy of the GNU Lesser General Public
+dnl  License along with this library; if not, write to the Free Software
+dnl  Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
+
+include(config.m4)
+include(ASM_SRCDIR/ppc.m4)
+
+define(`round',`
+	lwz r9,$3(r3)
+	xor $1,$1,r9
+	rlwinm r9,$1,10,22,29
+	rlwinm r10,$1,18,22,29
+	lwzx r9,r9,r28
+	lwzx r10,r10,r29
+	rlwinm r11,$1,26,22,29
+	add r9,r9,r10
+	lwzx r11,r11,r30
+	rlwinm r12,$1,2,22,29
+	xor r9,r9,r11
+	lwzx r12,r12,r31
+	add r9,r9,r12
+	xor $2,$2,r9
+')
+
+define(`eblock',`
+	round(r7,r8,0)
+	round(r8,r7,4)
+	round(r7,r8,8)
+	round(r8,r7,12)
+	round(r7,r8,16)
+	round(r8,r7,20)
+	round(r7,r8,24)
+	round(r8,r7,28)
+	round(r7,r8,32)
+	round(r8,r7,36)
+	round(r7,r8,40)
+	round(r8,r7,44)
+	round(r7,r8,48)
+	round(r8,r7,52)
+	round(r7,r8,56)
+	round(r8,r7,60)
+	lwz r9,64(r3)
+	lwz r10,68(r3)
+	xor r7,r7,r9
+	xor r8,r8,r10
+')
+
+define(`dblock',`
+	round(r7,r8,68)
+	round(r8,r7,64)
+	round(r7,r8,60)
+	round(r8,r7,56)
+	round(r7,r8,52)
+	round(r8,r7,48)
+	round(r7,r8,44)
+	round(r8,r7,40)
+	round(r7,r8,36)
+	round(r8,r7,32)
+	round(r7,r8,28)
+	round(r8,r7,24)
+	round(r7,r8,20)
+	round(r8,r7,16)
+	round(r7,r8,12)
+	round(r8,r7,8)
+	lwz r9,4(r3)
+	lwz r10,0(r3)
+	xor r7,r7,r9
+	xor r8,r8,r10
+')
+
+
+C_FUNCTION_BEGIN(blowfishEncrypt)
+	la r1,-16(r1)
+	stmw r28,0(r1)
+
+	la r28,72(r3)
+	la r29,1096(r3)
+	la r30,2120(r3)
+	la r31,3144(r3)
+
+ifelse(ASM_BIGENDIAN,yes,`
+	lwz r7,0(r5)
+	lwz r8,4(r5)
+',`
+	li r0,0
+	lwbrx r7,r5,r0
+	li r0,4
+	lwbrx r7,r5,r0
+')
+	
+	eblock
+
+ifelse(ASM_BIGENDIAN,yes,`
+	stw r7,4(r4)
+	stw r8,0(r4)
+',`
+	li r0,4
+	stwbrx r7,r4,r0
+	li r0,0
+	stwbrx r7,r4,r0
+')
+
+	li r3,0
+	lmw r28,0(r1)
+	la r1,16(r1)
+	blr
+C_FUNCTION_END(blowfishEncrypt)
+
+
+C_FUNCTION_BEGIN(blowfishDecrypt)
+	la r1,-16(r1)
+	stmw r28,0(r1)
+
+	la r28,72(r3)
+	la r29,1096(r3)
+	la r30,2120(r3)
+	la r31,3144(r3)
+
+ifelse(ASM_BIGENDIAN,yes,`
+	lwz r7,0(r5)
+	lwz r8,4(r5)
+',`
+	li r0,0
+	lwbrx r7,r5,r0
+	li r0,4
+	lwbrx r7,r5,r0
+')
+	
+	dblock
+
+ifelse(ASM_BIGENDIAN,yes,`
+	stw r7,4(r4)
+	stw r8,0(r4)
+',`
+	li r0,4
+	stwbrx r7,r4,r0
+	li r0,0
+	stwbrx r7,r4,r0
+')
+
+	li r3,0
+	lmw r28,0(r1)
+	la r1,16(r1)
+	blr
+C_FUNCTION_END(blowfishDecrypt)
diff --git a/beecrypt/gas/sha1opt.ia64.S b/beecrypt/gas/fips180opt.ia64.S
index 77a2a975a..a9c6edaa8 100644
--- a/beecrypt/gas/sha1opt.ia64.S
+++ b/beecrypt/gas/fips180opt.ia64.S
@@ -27,7 +27,7 @@
  *
  */
 
-#include "beecrypt.gas.h"
+#include "config.gas.h"
 
 #define saved_pfs	r14
 #define saved_lc	r15
@@ -38,13 +38,13 @@
 
 	.text
 
-#define K00	0x5a827999
-#define K20	0x6ed9eba1
-#define K40	0x8f1bbcdc
-#define K60	0xca62c1d6
+	.equ	K00,	0x5a827999
+	.equ	K20,	0x6ed9eba1
+	.equ	K40,	0x8f1bbcdc
+	.equ	K60,	0xca62c1d6
 
-#define PARAM_H		0
-#define PARAM_DATA	20
+	.equ    PARAM_H,        0
+	.equ    PARAM_DATA,     20
 
 /* for optimization, I have to see how I can parallellize the code
 
diff --git a/beecrypt/gas/ia64.m4 b/beecrypt/gas/ia64.m4
new file mode 100644
index 000000000..1ac0898d3
--- /dev/null
+++ b/beecrypt/gas/ia64.m4
@@ -0,0 +1,35 @@
+dnl  ia64.m4
+dnl
+dnl  Copyright (c) 2003 Bob Deblier
+dnl 
+dnl  Author: Bob Deblier <bob.deblier@pandora.be>
+dnl 
+dnl  This library is free software; you can redistribute it and/or
+dnl  modify it under the terms of the GNU Lesser General Public
+dnl  License as published by the Free Software Foundation; either
+dnl  version 2.1 of the License, or (at your option) any later version.
+dnl 
+dnl  This library 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 GNU
+dnl  Lesser General Public License for more details.
+dnl 
+dnl  You should have received a copy of the GNU Lesser General Public
+dnl  License along with this library; if not, write to the Free Software
+dnl  Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
+
+
+define(`saved_pfs',`r14')
+define(`saved_lc',`r15')
+
+ifelse(substr(ASM_OS,0,4),hpux,`
+undefine(`C_FUNCTION_BEGIN')
+define(C_FUNCTION_BEGIN,`
+	TEXTSEG
+	GLOBL SYMNAME($1)#
+	.proc SYMNAME($1)#
+SYMNAME($1):
+')
+',`
+	.explicit
+')
diff --git a/beecrypt/gas/m68k.m4 b/beecrypt/gas/m68k.m4
new file mode 100644
index 000000000..1c6bb6e1e
--- /dev/null
+++ b/beecrypt/gas/m68k.m4
@@ -0,0 +1,34 @@
+ifelse(REGISTERS_NEED_PERCENT,yes,`
+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)
+')
+ifelse(INSTRUCTIONS_NEED_DOT_SIZE_QUALIF,yes,`
+define(addal,adda.l)
+define(addl,add.l)
+define(addql,addq.l)
+define(addxl,addx.l)
+define(clrl,clr.l)
+define(lsll,lsl.l)
+define(movel,move.l)
+define(moveml,movem.l)
+define(moveal,movea.l)
+define(umull,umul.l)
+define(subl,sub.l)
+define(subql,subq.l)
+define(subxl,subx.l)
+')
diff --git a/beecrypt/gas/mp32opt.arm.S b/beecrypt/gas/mp32opt.arm.S
deleted file mode 100644
index 5908047ba..000000000
--- a/beecrypt/gas/mp32opt.arm.S
+++ /dev/null
@@ -1,180 +0,0 @@
-/*
- * mp32opt.arm.S
- *
- * Assembler optimized multiprecision integer routines for ARM processors
- *
- * Compile target is GNU Assembler
- *
- * Copyright (c) 2001, 2002 Virtual Unlimited B.V.
- *
- * Author: Bob Deblier <bob@virtualunlimited.com>
- *
- * This library is free software; you can redistribute it and/or
- * modify it under the terms of the GNU Lesser General Public
- * License as published by the Free Software Foundation; either
- * version 2.1 of the License, or (at your option) any later version.
- *
- * This library 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
- * Lesser General Public License for more details.
- *
- * You should have received a copy of the GNU Lesser General Public
- * License along with this library; if not, write to the Free Software
- * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
- *
- */
-
-#include "beecrypt.gas.h"
-
-	.file "mp32opt.arm.S"
-
-	.text
-
-/* ARM uses up to four registers for parameter passing */
-
-#if 0
-C_FUNCTION_BEGIN(mp32addw)
-LABEL(mp32addw)
-	/* r0 is xsize and must be at least one, r1 = xdata, r2 = y */
-	stmfd sp!, {r4, lr}
-	add r1, r1, r0, asl #2
-	mov r3, #0
-LOCAL(mp32addw_loop):
-	ldr r4, [r1, #-4]
-	adds r4, r4, r2
-	adc r2, r3, r3
-	str r4, [r1, #-4]!
-	subs r0, r0, #1
-	bne LOCAL(mp32addw_loop)
-	mov r0, r2
-	ldmfd sp!, {r4, pc}
-C_FUNCTION_END(mp32addw, LOCAL(mp32addw_size))
-
-
-C_FUNCTION_BEGIN(mp32subw)
-LABEL(mp32subw)
-	/* r0 is xsize and must be at least one, r1 = xdata, r2 = y */
-	stmfd sp!, {r4, lr}
-	add r1, r1, r0, asl #2
-	mov r3, #0
-LOCAL(mp32subw_loop):
-	ldr r4, [r1, #-4]
-	subs r4, r4, r2
-	adc r2, r3, r3
-	str r4, [r1, #-4]!
-	subs r0, r0, #1
-	bne LOCAL(mp32subw_loop)
-	mov r0, r2
-	ldmfd sp!, {r4, pc}
-C_FUNCTION_END(mp32subw, LOCAL(mp32subw_size))
-
-
-C_FUNCTION_BEGIN(mp32add)
-LABEL(mp32add)
-	/* r0 is size, r1 = xdata, r2 = ydata */
-	stmfd sp!, {r4, r5, lr}
-	/* copy cpsr to r5 and clear the carry bit */
-	mrs r5, cpsr
-	bic r5, r5, #0x20000000
-	/* adjust the addresses */
-	add r1, r1, r0, asl #2
-	add r2, r2, r0, asl #2
-LOCAL(mp32add_loop):
-	/* restore the carry bit */
-	msr cpsr_c, r5
-	ldr r3, [r1, #-4]!
-	ldr r4, [r2, #-4]!
-	adcs r3, r3, r2
-	str r3, [r1, #0]
-	/* save the carry bit */
-	mrs r5, cpsr
-	subs r0, r0, #1
-	bne LOCAL(mp32add_loop)
-
-	/* restore the carry bit */
-	msr cpsr_c, r5
-
-	/* set the result to the proper value */
-	adc r0, r0, r0
-	ldmfd sp!, {r4, r5, pc}
-C_FUNCTION_END(mp32add, LOCAL(mp32add_size))
-#endif
-
-
-C_FUNCTION_BEGIN(mp32setmul)
-LABEL(mp32setmul)
-	stmfd sp!, {r4, r5, lr}
-	/* adjust the addresses */
-	add r1, r1, r0, asl #2
-	add r2, r2, r0, asl #2
-	/* r3 is the multiplicand; r4 load from memory, r5 is scratch, ip is carry */
-	mov ip, #0
-LOCAL(mp32setmul_loop):
-	ldr r4, [r2, #-4]!
-	mov r5, #0
-	umlal ip, r5, r3, r4
-	str ip, [r1, #-4]!
-	mov ip, r5
-	subs r0, r0, #1
-	bne LOCAL(mp32setmul_loop)
-	/* return carry */
-	mov r0, ip
-	ldmfd sp!, {r4, r5, pc}
-C_FUNCTION_END(mp32setmul, LOCAL(mp32setmul_size))
-
-	
-C_FUNCTION_BEGIN(mp32addmul)
-LABEL(mp32addmul)
-	stmfd sp!, {r4, r5, r6, lr}
-	/* adjust the addresses */
-	add r1, r1, r0, asl #2
-	add r2, r2, r0, asl #2
-	/* r3 is the multiplicand; r4 & r5 load from memory, r6 is scratch, ip is carry */
-	mov ip, #0
-LOCAL(mp32addmul_loop):
-	ldr r4, [r2, #-4]!
-	ldr r5, [r1, #-4]
-	mov r6, #0
-	umlal ip, r6, r3, r4
-	adds r5, r5, ip
-	adc ip, r6, #0
-	str r5, [r1, #-4]!
-	subs r0, r0, #1
-	bne LOCAL(mp32addmul_loop)
-	/* return carry */
-	mov r0, ip
-	ldmfd sp!, {r4, r5, r6, pc}
-C_FUNCTION_END(mp32addmul, LOCAL(mp32addmul_size))
-
-
-#if 0
-/* this routine needs fixing; it causes a core dump for some reason */
-/* unfortunately the system I test this on has no debugger */
-C_FUNCTION_BEGIN(mp32addsqrtrc)
-LABEL(mp32addsqrtrc):
-	stmfd sp!, {r4, r5, r6, lr}
-	/* adjust the addresses */
-	add r1, r1, r0, asl #2
-	add r2, r2, r0, asl #2
-	/* r3 is a zero register, ip is the carry */
-	mov r3, #0
-	mov ip, #0
-LOCAL(mp32addsqrtrc_loop):
-	ldr r4, [r2, #-4]!
-	mov r6, #0
-	umlal ip, r6, r4, r4
-	ldr r5, [r1, #-4] /* lo word */
-	ldr r4, [r1, #-8] /* hi word */
-	adds r5, r5, ip
-	adcs r4, r4, r6
-	str r5, [r1, #-4]
-	str r4, [r1, #-8]!
-	adc ip, r3, #0 /* set carry */
-	subs r0, r0, #1
-	bne LOCAL(mp32addsqrtrc_loop)
-	/* return carry */
-	mov r0, ip
-	ldmfd sp!, {r4, r5, r6, pc}
-C_FUNCTION_END(mp32addsqrtrc, LOCAL(mp32addsqrtrc_size))
-#endif
diff --git a/beecrypt/gas/mp32opt.ia64.S b/beecrypt/gas/mp32opt.ia64.S
deleted file mode 100644
index 520ff4b52..000000000
--- a/beecrypt/gas/mp32opt.ia64.S
+++ /dev/null
@@ -1,260 +0,0 @@
-/*
- * mp32opt.ia64.S
- *
- * Assembler optimized multiprecision integer routines for ia64 (Intel Itanium)
- *
- * Compile target is GNU Assembler
- *
- * Copyright (c) 2000, 2001 Virtual Unlimited B.V.
- *
- * Author: Bob Deblier <bob@virtualunlimited.com>
- *
- * This library is free software; you can redistribute it and/or
- * modify it under the terms of the GNU Lesser General Public
- * License as published by the Free Software Foundation; either
- * version 2.1 of the License, or (at your option) any later version.
- *
- * This library 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
- * Lesser General Public License for more details.
- *
- * You should have received a copy of the GNU Lesser General Public
- * License along with this library; if not, write to the Free Software
- * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
- *
- */
-
-/*
- * I will need to implement 64 bit multiprecision assembler-optimized routines
- * before this platform can be tested adequately. The current 32 bit ones suffer
- * from loading into integer registers, conversion to floating point, doing the xma
- * and converting back to integer; 64 bit values can be loaded directly into
- * floating point registers, which should shave off a lot of cycles.
- */
-
-#include "beecrypt.gas.h"
-
-#define saved_pfs	r14
-#define saved_lc	r15
-
-#define size		r16
-#define dst			r17
-#define src			r18
-
-	.file	"mp32opt.ia64.S"
-
-	.text
-
-	.explicit
-
-	.align	32
-	.global	mpzero#
-	.proc	mpzero#
-
-mpzero:
-	alloc saved_pfs = ar.pfs,2,0,0,0
-	mov saved_lc = ar.lc
-	sub size = in0,r0,1;;
-	mov src = in1
-	mov ar.lc = size;;
-.L00:
-	st4 [src] = r0,4
-	br.ctop.sptk .L00
-	;;
-	mov ar.lc = saved_lc
-	mov ar.pfs = saved_pfs
-	br.ret.sptk b0
-	.endp	mpzero#
-
-
-	.align	32
-	.global	mpcopy#
-	.proc	mpcopy#
-
-mpcopy:
-	alloc saved_pfs = ar.pfs,3,5,0,8
-	mov saved_lc = ar.lc
-	sub size = in0,r0,1
-	mov dst = in1
-	mov src = in2;;
-	mov ar.lc = size
-	mov ar.ec = 2
-	mov pr.rot = (1 << 16);;
-.L01:
-	(p17) st4 [dst] = r33,4
-	(p16) ld4 r32 = [src],4;;
-	br.ctop.sptk .L01;;
-	mov ar.lc = saved_lc
-	mov ar.pfs = saved_pfs
-	br.ret.sptk b0
-	.endp	mpcopy#
-
-
-	.if		0
-	.align	32
-	.global	mpz
-	.type	mpz,@function
-
-mpz:
-	alloc r14 = ar.pfs,2,6,0,8
-	mov r15 = ar.lc
-	sub r16= in0,r0,1
-	mov r17 = in1
-	mov r18 = in2;;
-	mov r8 = 1
-	mov pr.rot = 1 << 16
-	mov ar.ec = 2
-	mov ar.lc = r16;;
-.L02:
-	(p16) ld4 r32 = [r18],4
-	(p18) cmp.eq p0,p32 = r34,r0
-	(p33) mov r8 = r0
-	(p33) br.exit
-	.endif
-
-
-	.align	32
-	.global	mpadd#
-	.proc	mpadd#
-
-mpadd:
-	alloc r14 = ar.pfs,3,0,0,0
-	mov r15 = ar.lc
-	# adjust size by -1
-	sub r16 = in0,r0,1
-	# clear carry
-	mov r8 = r0;;
-	# load addresses
-	shladd r17 = r16,2,in1
-	shladd r18 = r16,2,in2
-	# load loop count
-	mov ar.lc = r16;;
-.L20:
-	ld4 r20 = [r18],-4
-	ld4 r19 = [r17]
-	tbit.z p1,p2 = r8,32;;
-	(p1) add r8 = r19,r20
-	(p2) add r8 = r19,r20,1;;
-	st4 [r17] = r8,-4
-	br.cloop.sptk .L20;;
-	extr.u r8 = r8,32,1
-	mov ar.lc = r15
-	mov ar.pfs = r14
-	br.ret.sptk b0
-	.endp	mpadd#
-
-
-	.align	32
-	.global	mpsub#
-	.proc	mpsub#
-
-mpsub:
-	alloc r14 = ar.pfs,3,0,0,0
-	mov r15 = ar.lc
-	# adjust size by -1
-	sub r16 = in0,r0,1
-	# clear carry
-	mov r8 = r0;;
-	# load addresses
-	shladd r17 = r16,2,in1
-	shladd r18 = r16,2,in2
-	# load loop count
-	mov ar.lc = r16;;
-.L30:
-	ld4 r20 = [r18],-4
-	ld4 r19 = [r17]
-	tbit.z p1,p2 = r8,32;;
-	(p1) sub r8 = r19,r20
-	(p2) sub r8 = r19,r20,1;;
-	st4 [r17] = r8,-4
-	br.cloop.sptk .L30;;
-	extr.u r8 = r8,32,1
-	mov ar.lc = r15
-	mov ar.pfs = r14
-	br.ret.sptk b0
-	.endp	mpsub#
-
-
-	.if		0
-
-	.align	32
-	.global	mpsetmul#
-	.proc	mpsetmul#
-
-mpsetmul:
-	alloc r14 = ar.pfs,4,0,0,0
-	mov r15 = ar.lc
-	# load mul
-	setf.sig f96 = in3
-	# adjust size by -1
-	sub r16 = in0,r0,1
-	# clear carry
-	mov r8 = r0;;
-	# adjust addresses
-	shladd r17 = r16,2,in1
-	shladd r18 = r16,2,in2
-	# load loop count
-	mov ar.lc = r16;;
-.L40:
-	ld4 r19 = [r18],-4;;
-	setf.sig f98 = r8
-	setf.sig f97 = r19;;
-	# multiplication can only be done in f registers, but we do have a multiply-add
-	xma.l f98 = f96,f97,f98;;
-	getf.sig r8 = f98;;
-	st4 [r17] = r8,-4
-	shr.u r8 = r8,32
-	br.cloop.sptk .L40;;
-	mov ar.lc = r15
-	mov ar.pfs = r14
-	br.ret.sptk b0
-	.endp	mpsetmul#
-
-
-	.align	32
-	.global	mpaddmul#
-	.proc	mpaddmul#
-
-mpaddmul:
-	alloc saved_pfs = ar.pfs,4,0,0,0
-	mov saved_lc = ar.lc
-	# load mul
-	setf.sig f96 = in3
-	# adjust size by -1
-	sub size = in0,r0,1
-	# clear carry
-	mov r8 = r0;;
-	# adjust addresses
-	shladd dst = size,2,in1
-	shladd src = size,2,in2
-	# load loop count
-	mov ar.lc = r16;;
-.L50:
-	ld4 r19 = [dst]
-	ld4 r20 = [dst],-4;;
-	setf.sig f98 = r8
-	setf.sig f97 = r20;;
-	# multiplication can only be done in f registers, but we do have a multiply-add
-	xma.l f98 = f96,f97,f98;;
-	getf.sig r8 = f98;;
-	add r8 = r8,r19;;
-	st4 [r17] = r8,-4
-	shr.u r8 = r8,32
-	br.cloop.sptk .L50;;
-	mov ar.lc = r15
-	mov ar.pfs = r14
-	br.ret.sptk b0
-	.endp	mpaddmul#
-
-	.endif
-
-
-	.if 0
-	.align	16
-	.global	mpaddsqrtrc#
-	.proc	mpaddsqrtrc#
-
-mpaddsqrtrc:
-	.endp	mpaddsqrtrc#
-	.endif
diff --git a/beecrypt/gas/mp32opt.sparcv8.S b/beecrypt/gas/mp32opt.sparcv8.S
deleted file mode 100644
index 09a94f0d7..000000000
--- a/beecrypt/gas/mp32opt.sparcv8.S
+++ /dev/null
@@ -1,114 +0,0 @@
-/*
- * mp32opt.sparcv8.S
- *
- * Assembler optimized multiprecision integer routines for Sparc v8
- *
- * Compile target is GNU Assembler, Sun Solaris Assembler
- *
- * Copyright (c) 2001 Virtual Unlimited B.V.
- *
- * Author: Bob Deblier <bob@virtualunlimited.com>
- *
- * This library is free software; you can redistribute it and/or
- * modify it under the terms of the GNU Lesser General Public
- * License as published by the Free Software Foundation; either
- * version 2.1 of the License, or (at your option) any later version.
- *
- * This library 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
- * Lesser General Public License for more details.
- *
- * You should have received a copy of the GNU Lesser General Public
- * License along with this library; if not, write to the Free Software
- * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
- *
- */
-
-#include "beecrypt.gas.h"
-
-	.file "mp32opt.sparcv8.S"
-
-	.text
-
-C_FUNCTION_BEGIN(mpsetmul)
-LABEL(mpsetmul)
-
-	.register %g2,#scratch
-
-	sll %o0,2,%g1
-	dec 4,%o2
-	clr %o0
-LOCAL(mpsetmul_loop):
-	ld [%o2+%g1],%g2
-	umul %o3,%g2,%g2
-	rd %y,%g3
-	addcc %o0,%g2,%g2
-	addx %g0,%g3,%o0
-	deccc 4,%g1
-	bnz LOCAL(mpsetmul_loop)
-	st %g2,[%o1+%g1]
-	retl
-	nop
-C_FUNCTION_END(mpsetmul, LOCAL(mpsetmul_size))
-
-
-C_FUNCTION_BEGIN(mpaddmul)
-LABEL(mpaddmul)
-
-	.register %g2,#scratch
-
-	sll %o0,2,%g1
-	mov %o1,%o4
-	dec 4,%o1
-	dec 4,%o2
-	clr %o0
-LOCAL(mpaddmul_loop):
-	ld [%o2+%g1],%g2
-	ld [%o1+%g1],%g3
-	umul %o3,%g2,%g2
-	rd %y,%g4
-	addcc %o0,%g2,%g2
-	addx %g0,%g4,%g4
-	addcc %g2,%g3,%g2
-	addx %g0,%g4,%o0
-	deccc 4,%g1
-	bnz LOCAL(mpaddmul_loop)
-	st %g2,[%o4+%g1]
-	retl
-	nop
-C_FUNCTION_END(mpaddmul, LOCAL(mpaddmul_size))
-
-
-C_FUNCTION_BEGIN(mpaddsqrtrc)
-LABEL(mpaddsqrtrc)
-
-        .register %g2,#scratch
-        .register %g3,#scratch
-
-        sll %o0,2,%g1
-        add %o1,%g1,%o1
-		dec 4,%o2
-        add %o1,%g1,%o1
-		dec 8,%o1
-        clr %o0
-LOCAL(mpaddsqrtrc_loop):
-        ld [%o2+%g1],%g2
-        ldd [%o1],%o4
-        umul %g2,%g2,%g3
-        rd %y,%g2
-		/* first addition */
-        addcc %o5,%g3,%o5
-        addxcc %o4,%g2,%o4
-        addx %g0,%g0,%o3
-		/* second addition */
-        addcc %o5,%o0,%o5
-        addxcc %o4,%g0,%o4
-        addx %o3,%g0,%o0
-        std %o4,[%o1]
-        deccc 4,%g1
-        bnz LOCAL(mpaddsqrtrc_loop)
-        sub %o1,8,%o1
-        retl
-        nop
-C_FUNCTION_END(mpaddsqrtrc, LOCAL(mpaddsqrtrc_size))
diff --git a/beecrypt/gas/mp64opt.ia64.S b/beecrypt/gas/mp64opt.ia64.S
deleted file mode 100644
index 581bf4d30..000000000
--- a/beecrypt/gas/mp64opt.ia64.S
+++ /dev/null
@@ -1,322 +0,0 @@
-/*
- * mp64opt.ia64.S
- *
- * Assembler optimized multiprecision integer routines for ia64 (Intel Itanium)
- *
- * Compile target is GNU Assembler
- *
- * Copyright (c) 2000, 2001 Virtual Unlimited B.V.
- *
- * Author: Bob Deblier <bob@virtualunlimited.com>
- *
- * This library is free software; you can redistribute it and/or
- * modify it under the terms of the GNU Lesser General Public
- * License as published by the Free Software Foundation; either
- * version 2.1 of the License, or (at your option) any later version.
- *
- * This library 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
- * Lesser General Public License for more details.
- *
- * You should have received a copy of the GNU Lesser General Public
- * License along with this library; if not, write to the Free Software
- * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
- *
- */
-
-#include "beecrypt.gas.h"
-
-#define saved_pfs	r14
-#define	saved_lc	r15
-
-#define size		r16
-#define dst			r17
-#define src			r18
-#define alt			r19
-
-	.text
-
-	.explicit
-
-/* functions to add, in order of priority:
- * mp64addsqrtrc
- * mp64neg -> can vectorize
- * mp64multwo -> can vectorize
- * mp32divtwo -> ..
- * mp64fill -> easy
- * mp64z -> vectorizable with br.wtop
- * mp64nz -> vectorizable with br.wtop
- * mp64eq -> ..
- * mp64eqx -> ..
- * mp64ne -> ..
- * mp64nex -> ..
- * mp64gt -> ..
- * mp64gtx -> ..
- * mp64lt -> ..
- * mp64ltx -> ..
- * mp64ge -> substitute with mp64lt with swap of parameters
- * mp64gex -> .. mp64ltx
- * mp64le -> .. mp64gt
- * mp64lex -> .. mp64gtx
- * mp64isone -> vectorizable with br.wtop
- * mp64istwo -> ..
- * mp64leone -> ..
- * mp64size -> ..
-
-/* mp64zero works */
-C_FUNCTION_BEGIN(mp64zero)
-	alloc saved_pfs = ar.pfs,2,0,0,0
-	mov saved_lc = ar.lc
-	sub size = in0,r0,1;;
-	mov src = in1
-	mov ar.lc = size;;
-
-.Lmp64zero_loop:
-	st8 [src] = r0,8
-	br.ctop.sptk .Lmp64zero_loop;;
-
-	mov ar.lc = saved_lc
-	mov ar.pfs = saved_pfs
-	br.ret.sptk b0
-C_FUNCTION_END(mp64zero)
-
-
-/* mp64copy works */
-C_FUNCTION_BEGIN(mp64copy)
-	alloc saved_pfs = ar.pfs,3,5,0,8
-	mov saved_lc = ar.lc
-	sub size = in0,r0,1;;
-	mov dst = in1
-	mov src = in2
-	/* prepare loop */
-	mov ar.lc = size
-	mov ar.ec = 2
-	mov pr.rot = (1 << 16);;
-
-.Lmp64copy_loop:
-	(p17) st8 [dst] = r33,-8
-	(p16) ld8 r32 = [src],-8;;
-	br.ctop.sptk .Lmp64copy_loop;;
-
-	mov ar.lc = saved_lc
-	mov ar.pfs = saved_pfs
-	br.ret.sptk b0
-C_FUNCTION_END(mp64copy)
-
-
-#if 0
-/* mp64z is in development */
-C_FUNCTION_BEGIN(mp64z)
-	alloc saved_pfs = ar.pfs,2,6,0,8
-	mov saved_lc = ar.lc
-	sub size = in0,r0,1;;
-
-	mov ret0 = 1
-	mov src = in1
-
-	mov ar.lc = size
-	mov ar.ec = 2
-	mov pr.rot = ((1 << 16) | (1 << 20));;
-
-.Lmp64z_loop:
-	(p16) ld8 r32 = [src],8
-	(p17) cmp.ne p1,p0 = r33,r0
-	(p1) br.exit.dpnt .Lmp64z_exit;;
-	br.ctop.dptk .Lmp64z_loop;;
-.Lmp64z_exit:
-	(p1) mov ret0 = r0
-
-	mov ar.lc = saved_lc
-	mov ar.pfs = saved_pfs
-	br.ret.sptk b0
-C_FUNCTION_END(mp64z)
-#endif
-
-
-/* mp64add works */
-C_FUNCTION_BEGIN(mp64add)
-	alloc saved_pfs = ar.pfs,3,5,0,8
-	mov saved_lc = ar.lc
-	sub size = in0,r0,1;;
-
-	/* adjust addresses */
-	shladd dst = size,3,in1
-	shladd src = size,3,in2
-	shladd alt = size,3,in1
-
-	/* prepare modulo-scheduled loop */
-	mov ar.lc = size
-	mov ar.ec = 3 
-	mov pr.rot = ((1 << 16) | (1 << 19));;
-
-.Lmp64add_loop:
-	(p16) ld8 r32 = [src],-8
-	(p16) ld8 r35 = [alt],-8
-	(p20) add r36 = r33,r36		/* no carry add */
-	(p22) add r36 = r33,r36,1	/* carry add */
-	;;
-	(p20) cmp.leu p19,p21 = r33,r36	/* no previous carry */
-	(p22) cmp.ltu p19,p21 = r33,r36	/* previous carry */
-	(p18) st8 [dst] = r37,-8
-	br.ctop.dptk .Lmp64add_loop;;
-
-	/* return carry */
-	(p21) add ret0 = r0,r0
-	(p23) add ret0 = r0,r0,1
-	;;
-	mov ar.lc = saved_lc
-	mov ar.pfs = saved_pfs
-	br.ret.sptk b0
-C_FUNCTION_END(mp64add)
-
-
-/* mp64sub is in development */
-C_FUNCTION_BEGIN(mp64sub)
-	alloc saved_pfs = ar.pfs,3,5,0,8
-	mov saved_lc = ar.lc
-	sub size = in0,r0,1;;
-
-	/* adjust addresses */
-	shladd dst = size,3,in1
-	shladd src = size,3,in2
-	shladd alt = size,3,in1
-
-	/* prepare modulo-scheduled loop */
-	mov ar.lc = size
-	mov ar.ec = 3 
-	mov pr.rot = ((1 << 16) | (1 << 19));;
-
-.Lmp64sub_loop:
-	(p16) ld8 r32 = [src],-8
-	(p16) ld8 r35 = [alt],-8
-	(p20) sub r36 = r33,r36		/* no carry sub */
-	(p22) sub r36 = r33,r36,1	/* carry sub */
-	;;
-	(p20) cmp.geu p19,p21 = r33,r36	/* no previous carry */
-	(p22) cmp.gtu p19,p21 = r33,r36	/* previous carry */
-	(p18) st8 [dst] = r37,-8
-	br.ctop.dptk .Lmp64sub_loop;;
-
-	/* return carry */
-	(p21) add ret0 = r0,r0
-	(p23) add ret0 = r0,r0,1
-	;;
-	mov ar.lc = saved_lc
-	mov ar.pfs = saved_pfs
-	br.ret.sptk b0
-C_FUNCTION_END(mp64sub)
-
-
-/* mp64setmul works */
-C_FUNCTION_BEGIN(mp64setmul)
-	alloc saved_pfs = ar.pfs,4,4,0,8
-	mov saved_lc = ar.lc
-
-	setf.sig f6 = in3	/* the multiplier */
-	setf.sig f7 = r0 	/* the carry */
-	sub size = in0,r0,1;;
-
-	/* adjust addresses */
-	shladd dst = size,3,in1
-	shladd src = size,3,in2
-
-	/* prepare modulo-scheduled loop */
-	mov ar.lc = size
-	mov ar.ec = 3
-	mov pr.rot = (1 << 16);;
-
-.Lmp64setmul_loop:
-	(p16) ldf8 f36 = [src],-8
-	(p18) stf8 [dst] = f33,-8
-	(p17) xma.lu f32 = f6,f37,f7
-	(p17) xma.hu f7  = f6,f37,f7;;
-	br.ctop.dptk .Lmp64setmul_loop;;
-
-	/* return carry */
-	getf.sig ret0 = f7;;
-
-	mov ar.lc = saved_lc
-	mov ar.pfs = saved_pfs
-	br.ret.sptk b0
-C_FUNCTION_END(mp64setmul)
-
-
-/* mp64addmul needs fixing */
-C_FUNCTION_BEGIN(mp64addmul)
-	alloc saved_pfs = ar.pfs,4,12,0,16
-	mov saved_lc = ar.lc
-
-	sub size = in0,r0,1;;
-	setf.sig f6 = in3	/* the multiplier */
-
-	/* adjust addresses */
-	shladd dst = size,3,in1
-	shladd src = size,3,in2
-	shladd alt = size,3,in1;;
-
-	/* prepare the rotate-in carry */
-	mov r32 = r0		
-
-	/* prepare modulo-scheduled loop */
-	mov ar.lc = size
-	mov ar.ec = 5
-	mov pr.rot = ((1 << 16) | (1 << 21));
-
-.Lmp64addmul_loop:
-	(p18) getf.sig r33 = f34	/* hi 64 bit word */
-	(p24) add r38 = r35,r38
-	(p17) xma.lu f37 = f6,f41,f45
-	(p18) getf.sig r37 = f38	/* lo 64 bit word */
-	(p26) add r38 = r35,r38,1
-	(p17) xma.hu f33 = f6,f41,f45
-	(p16) ldf8 f40 = [src],-8
-	(p16) ldf8 f44 = [alt],-8
-	;;
-	/* set carry from this operation */
-	(p24) cmp.leu p23,p25 = r35,r38
-	(p26) cmp.ltu p23,p25 = r35,r38
-	(p20) st8 [dst] = r39,-8
-	br.ctop.dptk .Lmp64addmul_loop;;
-
-	/* return carry */
-	(p25) add ret0 = r36,r0
-	(p27) add ret0 = r36,r0,1
-
-	mov ar.lc = saved_lc
-	mov ar.pfs = saved_pfs
-	br.ret.sptk b0
-C_FUNCTION_END(mp64addmul)
-
-/* mp64addsqrtrc will be a little more challenging */
-
-/* the primary loop will look like this:
-
-.Lmp64addsqrtrc_loop:
-	/* stage 1 */
-	(p16) ldf8 to_square
-	(p16) ld8 lo_to_add
-	(p16) ld8 hi_to_add
-	/* stage 2 */
-	(p17) xma.lu to_square,to_square,carry
-	(p17) xma.hu to_square,to_square,carry
-	/* stage 3 */
-	(p18) getf lo xma
-	(p18) getf hi xma
-	/* stage 4 */
-	(p?) add lo no carry 
-	(p?) add lo carry
-	/* stage 5 */
-    (p?+1) add hi no carry
-	(p?+1) add hi carry
-	;;
-	/* also stage 4 */
-	(p?) cmp lo for carry
-	(p?) cmp lo for carry
-	/* also stage 5 */
-	(p?+1) cmp hi for carry
-	(p?+1) cmp hi for carry
-	st8 lo
-	st8 hi
-	br.ctop
-*/
diff --git a/beecrypt/gas/mpopt.alpha.m4 b/beecrypt/gas/mpopt.alpha.m4
new file mode 100644
index 000000000..55d4b5266
--- /dev/null
+++ b/beecrypt/gas/mpopt.alpha.m4
@@ -0,0 +1,159 @@
+dnl  mpopt.alpha.m4
+dnl
+dnl  Copyright (c) 2003 Bob Deblier
+dnl
+dnl  Author: Bob Deblier <bob.deblier@pandora.be>
+dnl
+dnl  This library is free software; you can redistribute it and/or
+dnl  modify it under the terms of the GNU Lesser General Public
+dnl  License as published by the Free Software Foundation; either
+dnl  version 2.1 of the License, or (at your option) any later version.
+dnl
+dnl  This library 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 GNU
+dnl  Lesser General Public License for more details.
+dnl
+dnl  You should have received a copy of the GNU Lesser General Public
+dnl  License along with this library; if not, write to the Free Software
+dnl  Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
+
+include(config.m4)
+include(ASM_SRCDIR/alpha.m4)
+
+
+C_FUNCTION_BEGIN(mpadd)
+	subq `$'16,1,`$'16
+	s8addq `$'16,0,`$'1
+	addq `$'17,`$'1,`$'17
+	addq `$'18,`$'1,`$'18
+	mov `$31',`$'0
+
+	.align 4
+LOCAL(mpadd_loop):
+	ldq `$'1,0(`$'17)
+	ldq `$'2,0(`$'18)
+	addq `$'1,`$'0,`$'3
+	cmpult `$'3,`$'1,`$'0
+	addq `$3',`$'2,`$'1
+	cmpult `$'1,`$'3,`$'2
+	stq `$'1,0(`$'17)
+	or `$'2,`$'0,`$'0
+	subq `$'16,1,`$'16
+	subq `$'17,8,`$'17
+	subq `$'18,8,`$'18
+	bge `$'16,LOCAL(mpadd_loop)
+	ret `$'31,(`$'26),1
+C_FUNCTION_END(mpadd)
+
+
+C_FUNCTION_BEGIN(mpsub)
+	subq `$'16,1,`$'16
+	s8addq `$'16,0,`$'1
+	addq `$'17,`$'1,`$'17
+	addq `$'18,`$'1,`$'18
+	mov `$31',`$'0
+
+	.align 4
+LOCAL(mpsub_loop):
+	ldq `$'1,0(`$'17)
+	ldq `$'2,0(`$'18)
+	subq `$'1,`$'0,`$'3
+	cmpult `$'1,`$'3,`$'0
+	subq `$'3,`$'2,`$'1
+	cmpult `$'3,`$'1,`$'2
+	stq `$'1,0(`$'17)
+	or `$'2,`$'0,`$'0
+	subq `$'16,1,`$'16
+	subq `$'17,8,`$'17
+	subq `$'18,8,`$'18
+	bge `$'16,LOCAL(mpsub_loop)
+	ret `$'31,(`$'26),1
+C_FUNCTION_END(mpsub)
+
+
+C_FUNCTION_BEGIN(mpsetmul)
+	subq `$'16,1,`$'16
+	s8addq `$'16,0,`$'1
+	addq `$'17,`$'1,`$'17
+	addq `$'18,`$'1,`$'18
+	mov `$31',`$'0
+
+	.align 4
+LOCAL(mpsetmul_loop):
+	ldq `$1',0(`$'18)
+	mulq `$'19,`$'1,`$'2
+	umulh `$'19,`$'1,`$'3
+	addq `$'2,`$'0,`$'2
+	cmpult `$'2,`$'0,`$'0
+	stq `$'2,0(`$'17)
+	addq `$'3,`$'0,`$'0
+	subq `$'16,1,`$'16
+	subq `$'17,8,`$'17
+	subq `$'18,8,`$'18
+	bge `$'16,LOCAL(mpsetmul_loop)
+	ret `$'31,(`$'26),1
+C_FUNCTION_END(mpsetmul)
+
+
+C_FUNCTION_BEGIN(mpaddmul)
+	subq `$'16,1,`$'16
+	s8addq `$'16,0,`$'1
+	addq `$'17,`$'1,`$'17
+	addq `$'18,`$'1,`$'18
+	mov `$31',`$'0
+
+	.align 4
+LOCAL(mpaddmul_loop):
+	ldq `$'1,0(`$'17)
+	ldq `$'2,0(`$'18)
+	mulq `$'19,`$'2,`$'3
+	umulh `$'19,`$'2,`$'4
+	addq `$'3,`$'0,`$'3
+	cmpult `$'3,`$'0,`$'0
+	addq `$'4,`$'0,`$'4
+	addq `$'3,`$'1,`$'3
+	cmpult `$'3,`$'1,`$'0
+	addq `$'4,`$'0,`$'0
+	stq `$'3,0(`$'17)
+	subq `$'16,1,`$'16
+	subq `$'17,8,`$'17
+	subq `$'18,8,`$'18
+	bge `$'16,LOCAL(mpaddmul_loop)
+	ret `$'31,(`$'26),1
+C_FUNCTION_END(mpaddmul)
+
+
+C_FUNCTION_BEGIN(mpaddsqrtrc)
+	subq `$'16,1,`$'16
+	s8addq `$'16,0,`$'1
+	addq `$'17,`$'1,`$'17
+	addq `$'17,`$'1,`$'17
+	addq `$'18,`$'1,`$'18
+	mov `$31',`$'0
+
+	.align 4
+LOCAL(mpaddsqrtrc_loop):
+	ldq `$'1,0(`$'18)
+	mulq `$1',`$1',`$'2
+	umulh `$1',`$1',`$'1
+	addq `$'2,`$'0,`$'3
+	cmpult `$3',`$'2,`$'0
+	ldq `$'2,8(`$'17)
+	addq `$'1,`$'0,`$'1
+	addq `$'3,`$'2,`$'4
+	cmpult `$'4,`$'3,`$'0
+	ldq `$'3,0(`$'17)
+	addq `$'1,`$'0,`$'2
+	cmpult `$2',`$'1,`$'0
+	stq `$'4,8(`$'17)
+	addq `$'2,`$'3,`$'1
+	cmpult `$'1,`$'2,`$2'
+	stq `$'1,0(`$'17)
+	addq `$'2,`$'0,`$'0
+	subq `$'16,1,`$'16
+	subq `$'17,16,`$'17
+	subq `$'18,8,`$'18
+	bge `$'16,LOCAL(mpaddmul_loop)
+	ret `$'31,(`$'26),1
+C_FUNCTION_END(mpaddsqrtrc)
diff --git a/beecrypt/gas/mpopt.arm.m4 b/beecrypt/gas/mpopt.arm.m4
new file mode 100644
index 000000000..a7dc677ee
--- /dev/null
+++ b/beecrypt/gas/mpopt.arm.m4
@@ -0,0 +1,83 @@
+dnl  mpopt.arm.m4
+dnl
+dnl  Copyright (c) 2003 Bob Deblier
+dnl 
+dnl  Author: Bob Deblier <bob.deblier@pandora.be>
+dnl 
+dnl  This library is free software; you can redistribute it and/or
+dnl  modify it under the terms of the GNU Lesser General Public
+dnl  License as published by the Free Software Foundation; either
+dnl  version 2.1 of the License, or (at your option) any later version.
+dnl 
+dnl  This library 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 GNU
+dnl  Lesser General Public License for more details.
+dnl 
+dnl  You should have received a copy of the GNU Lesser General Public
+dnl  License along with this library; if not, write to the Free Software
+dnl  Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
+ 
+include(config.m4)
+
+
+C_FUNCTION_BEGIN(mpsetmul)
+	stmfd sp!, {r4, r5, lr}
+	add r1, r1, r0, asl #2
+	add r2, r2, r0, asl #2
+	mov ip, #0
+LOCAL(mpsetmul_loop):
+	ldr r4, [r2, #-4]!
+	mov r5, #0
+	umlal ip, r5, r3, r4
+	str ip, [r1, #-4]!
+	mov ip, r5
+	subs r0, r0, #1
+	bne LOCAL(mpsetmul_loop)
+	mov r0, ip
+	ldmfd sp!, {r4, r5, pc}
+C_FUNCTION_END(mpsetmul)
+
+	
+C_FUNCTION_BEGIN(mpaddmul)
+	stmfd sp!, {r4, r5, r6, lr}
+	add r1, r1, r0, asl #2
+	add r2, r2, r0, asl #2
+	mov ip, #0
+LOCAL(mpaddmul_loop):
+	ldr r4, [r2, #-4]!
+	ldr r5, [r1, #-4]
+	mov r6, #0
+	umlal ip, r6, r3, r4
+	adds r5, r5, ip
+	adc ip, r6, #0
+	str r5, [r1, #-4]!
+	subs r0, r0, #1
+	bne LOCAL(mpaddmul_loop)
+	mov r0, ip
+	ldmfd sp!, {r4, r5, r6, pc}
+C_FUNCTION_END(mpaddmul)
+
+
+C_FUNCTION_BEGIN(mpaddsqrtrc)
+	stmfd sp!, {r4, r5, r6, lr}
+	add r1, r1, r0, asl #3
+	add r2, r2, r0, asl #2
+	mov r3, #0
+	mov ip, #0
+LOCAL(mpaddsqrtrc_loop):
+	ldr r4, [r2, #-4]!
+	mov r6, #0
+	umlal ip, r6, r4, r4
+	ldr r5, [r1, #-4]
+	ldr r4, [r1, #-8]
+	adds r5, r5, ip
+	adcs r4, r4, r6
+	str r5, [r1, #-4]
+	str r4, [r1, #-8]!
+	adc ip, r3, #0
+	subs r0, r0, #1
+	bne LOCAL(mpaddsqrtrc_loop)
+	mov r0, ip
+	ldmfd sp!, {r4, r5, r6, pc}
+C_FUNCTION_END(mpaddsqrtrc)
diff --git a/beecrypt/gas/mpopt.ia64.m4 b/beecrypt/gas/mpopt.ia64.m4
new file mode 100644
index 000000000..8486fe2e5
--- /dev/null
+++ b/beecrypt/gas/mpopt.ia64.m4
@@ -0,0 +1,187 @@
+dnl  mpopt.ia64.m4
+dnl
+dnl  Copyright (c) 2003 Bob Deblier
+dnl 
+dnl  Author: Bob Deblier <bob.deblier@pandora.be>
+dnl 
+dnl  This library is free software; you can redistribute it and/or
+dnl  modify it under the terms of the GNU Lesser General Public
+dnl  License as published by the Free Software Foundation; either
+dnl  version 2.1 of the License, or (at your option) any later version.
+dnl 
+dnl  This library 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 GNU
+dnl  Lesser General Public License for more details.
+dnl 
+dnl  You should have received a copy of the GNU Lesser General Public
+dnl  License along with this library; if not, write to the Free Software
+dnl  Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
+
+include(config.m4)
+include(ASM_SRCDIR/ia64.m4)
+
+define(`size',`r16')
+define(`dst',`r17')
+define(`src',`r18')
+define(`alt',`r19')
+
+
+C_FUNCTION_BEGIN(mpadd)
+	alloc saved_pfs = ar.pfs,3,5,0,8
+	mov saved_lc = ar.lc
+	sub size = in0,r0,1;;
+
+dnl	adjust addresses
+	shladd dst = size,3,in1
+	shladd src = size,3,in2
+	shladd alt = size,3,in1
+
+dnl	prepare modulo-scheduled loop
+	mov ar.lc = size
+	mov ar.ec = 2 
+	mov pr.rot = ((1 << 16) | (1 << 19));;
+
+LOCAL(mpadd_loop):
+	(p16) ld8 r32 = [alt],-8
+	(p16) ld8 r35 = [src],-8
+	(p20) add r36 = r33,r36
+	(p22) add r36 = r33,r36,1
+	;;
+	(p20) cmp.leu p19,p21 = r33,r36
+	(p22) cmp.ltu p19,p21 = r33,r36
+	(p18) st8 [dst] = r37,-8
+	br.ctop.dptk LOCAL(mpadd_loop);;
+
+dnl	loop epilogue: final store
+	(p18) st8 [dst] = r37,-8
+
+dnl	return carry
+	(p20) add ret0 = r0,r0
+	(p22) add ret0 = r0,r0,1
+	;;
+	mov ar.lc = saved_lc
+	mov ar.pfs = saved_pfs
+	br.ret.sptk b0
+C_FUNCTION_END(mpadd)
+
+
+C_FUNCTION_BEGIN(mpsub)
+	alloc saved_pfs = ar.pfs,3,5,0,8
+	mov saved_lc = ar.lc
+	sub size = in0,r0,1;;
+
+dnl	adjust addresses
+	shladd dst = size,3,in1
+	shladd src = size,3,in2
+	shladd alt = size,3,in1
+
+dnl	prepare modulo-scheduled loop
+	mov ar.lc = size
+	mov ar.ec = 2 
+	mov pr.rot = ((1 << 16) | (1 << 19));;
+
+LOCAL(mpsub_loop):
+	(p16) ld8 r32 = [alt],-8
+	(p16) ld8 r35 = [src],-8
+	(p20) sub r36 = r33,r36
+	(p22) sub r36 = r33,r36,1
+	;;
+	(p20) cmp.geu p19,p21 = r33,r36
+	(p22) cmp.gtu p19,p21 = r33,r36
+	(p18) st8 [dst] = r37,-8
+	br.ctop.dptk LOCAL(mpsub_loop);;
+
+dnl	loop epilogue: final store
+	(p18) st8 [dst] = r37,-8
+
+dnl	return carry
+	(p20) add ret0 = r0,r0
+	(p22) add ret0 = r0,r0,1
+	;;
+	mov ar.lc = saved_lc
+	mov ar.pfs = saved_pfs
+	br.ret.sptk b0
+C_FUNCTION_END(mpsub)
+
+
+C_FUNCTION_BEGIN(mpsetmul)
+	alloc saved_pfs = ar.pfs,4,4,0,8
+	mov saved_lc = ar.lc
+
+	setf.sig f6 = in3
+	setf.sig f7 = r0
+	sub size = in0,r0,1;;
+
+dnl	adjust addresses
+	shladd dst = size,3,in1
+	shladd src = size,3,in2
+
+dnl	prepare modulo-scheduled loop
+	mov ar.lc = size
+	mov ar.ec = 3
+	mov pr.rot = (1 << 16);;
+
+LOCAL(mpsetmul_loop):
+	(p16) ldf8 f32 = [src],-8
+	(p18) stf8 [dst] = f35,-8
+	(p17) xma.lu f34 = f6,f33,f7
+	(p17) xma.hu f7  = f6,f33,f7;;
+	br.ctop.dptk LOCAL(mpsetmul_loop);;
+
+dnl	return carry
+	getf.sig ret0 = f7;;
+
+	mov ar.lc = saved_lc
+	mov ar.pfs = saved_pfs
+	br.ret.sptk b0
+C_FUNCTION_END(mpsetmul)
+
+
+C_FUNCTION_BEGIN(mpaddmul)
+	alloc saved_pfs = ar.pfs,4,4,0,8
+	mov saved_lc = ar.lc
+
+	setf.sig f6 = in3
+	sub size = in0,r0,1;;
+
+dnl	adjust addresses
+	shladd dst = size,3,in1
+	shladd src = size,3,in2
+	shladd alt = size,3,in1;;
+
+dnl	prepare the rotate-in carry
+	mov r32 = r0		
+
+dnl	prepare modulo-scheduled loop
+	mov ar.lc = size
+	mov ar.ec = 4
+	mov pr.rot = ((1 << 16) | (1 << 21));
+
+LOCAL(mpaddmul_loop):
+	(p18) getf.sig r37 = f35
+	(p24) add r35 = r38,r35
+	(p17) xma.lu f34 = f6,f33,f37
+	(p18) getf.sig r33 = f39
+	(p26) add r35 = r38,r35,1
+	(p17) xma.hu f38 = f6,f33,f37
+	(p16) ldf8 f32 = [src],-8
+	(p16) ldf8 f36 = [alt],-8
+	;;
+dnl	set carry from this operation
+	(p24) cmp.leu p23,p25 = r38,r35
+	(p26) cmp.ltu p23,p25 = r38,r35
+	(p20) st8 [dst] = r36,-8
+	br.ctop.dptk LOCAL(mpaddmul_loop);;
+
+dnl	loop epilogue: final store
+	(p20) st8 [dst] = r36,-8
+
+dnl	return carry
+	(p24) add ret0 = r35,r0
+	(p26) add ret0 = r35,r0,1
+
+	mov ar.lc = saved_lc
+	mov ar.pfs = saved_pfs
+	br.ret.sptk b0
+C_FUNCTION_END(mpaddmul)
diff --git a/beecrypt/gas/mpopt.m68k.m4 b/beecrypt/gas/mpopt.m68k.m4
new file mode 100644
index 000000000..0cb2d4ca3
--- /dev/null
+++ b/beecrypt/gas/mpopt.m68k.m4
@@ -0,0 +1,158 @@
+dnl  mpopt.m68k.m4
+dnl
+dnl  Copyright (c) 2003 Bob Deblier
+dnl 
+dnl  Author: Bob Deblier <bob.deblier@pandora.be>
+dnl 
+dnl  This library is free software; you can redistribute it and/or
+dnl  modify it under the terms of the GNU Lesser General Public
+dnl  License as published by the Free Software Foundation; either
+dnl  version 2.1 of the License, or (at your option) any later version.
+dnl 
+dnl  This library 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 GNU
+dnl  Lesser General Public License for more details.
+dnl 
+dnl  You should have received a copy of the GNU Lesser General Public
+dnl  License along with this library; if not, write to the Free Software
+dnl  Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
+ 
+include(config.m4)
+include(ASM_SRCDIR/m68k.m4)
+
+dnl works
+C_FUNCTION_BEGIN(mpadd)
+	move.l 4(%sp),%d0
+	movea.l 8(%sp),%a0
+	movea.l 12(%sp),%a1
+	move.l %d0,%d1
+	lsl.l #2,%d0
+	subq.l #1,%d1
+	adda.l %d0,%a0
+	adda.l %d0,%a1
+	clr %d0
+
+	.align 2
+LOCAL(mpadd_loop):
+	addx.l -(%a1),-(%a0)
+	dbf %d1,LOCAL(mpadd_loop)
+
+	addx.l %d0,%d0
+	rts
+C_FUNCTION_END(mpadd)
+
+dnl works
+C_FUNCTION_BEGIN(mpsub)
+	move.l 4(%sp),%d0
+	movea.l 8(%sp),%a0
+	movea.l 12(%sp),%a1
+	move.l %d0,%d1
+	lsl.l #2,%d0
+	subq.l #1,%d1
+	adda.l %d0,%a0
+	adda.l %d0,%a1
+	clr %d0
+
+	.align 2
+LOCAL(mpsub_loop):
+	subx.l -(%a1),-(%a0)
+	dbf %d1,LOCAL(mpsub_loop)
+
+	addx.l %d0,%d0
+	rts
+C_FUNCTION_END(mpsub)
+
+dnl works
+C_FUNCTION_BEGIN(mpsetmul)
+	movem.l %d2-%d5,-(%sp)
+	move.l 20(%sp),%d0
+	movea.l 24(%sp),%a0
+	movea.l 28(%sp),%a1
+	move.l 32(%sp),%d2
+	move.l %d0,%d5
+	lsl.l #2,%d0
+	subq.l #1,%d5
+	adda.l %d0,%a0
+	adda.l %d0,%a1
+	clr.l %d3
+	clr.l %d4
+
+	.align 2
+LOCAL(mpsetmul_loop):
+	move.l -(%a1),%d1
+	mulu.l %d2,%d0:%d1
+	add.l  %d3,%d1
+	addx.l %d4,%d0
+	move.l %d1,-(%a0)
+	move.l %d0,%d3
+	dbf %d5,LOCAL(mpsetmul_loop)
+
+	movem.l (%sp)+,%d2-%d5
+	rts
+C_FUNCTION_END(mpsetmul)
+
+dnl works
+C_FUNCTION_BEGIN(mpaddmul)
+	movem.l %d2-%d5,-(%sp)
+	move.l 20(%sp),%d0
+	movea.l 24(%sp),%a0
+	movea.l 28(%sp),%a1
+	move.l 32(%sp),%d2
+	move.l %d0,%d5
+	lsl.l #2,%d0
+	subq.l #1,%d5
+	adda.l %d0,%a0
+	adda.l %d0,%a1
+	clr.l %d3
+	clr.l %d4
+
+	.align 2
+LOCAL(mpaddmul_loop):
+	move.l -(%a1),%d1
+	mulu.l %d2,%d0:%d1
+	add.l %d3,%d1
+	addx.l %d4,%d0
+	add.l -(%a0),%d1
+	addx.l %d4,%d0
+	move.l %d1,(%a0)
+	move.l %d0,%d3
+	dbf %d5,LOCAL(mpaddmul_loop)
+
+	movem.l (%sp)+,%d2-%d5
+	rts
+C_FUNCTION_END(mpaddmul)
+
+
+C_FUNCTION_BEGIN(mpaddsqrtrc)
+	movem.l %d3-%d5,-(%sp)
+	move.l 16(%sp),%d0
+	movea.l 20(%sp),%a0
+	movea.l 24(%sp),%a1
+	move.l %d0,%d5
+	lsl.l #2,%d0
+	subq.l #1,%d5
+	adda.l %d0,%a0
+	adda.l %d0,%a0
+	adda.l %d0,%a1
+	clr.l %d3
+	clr.l %d4
+
+LOCAL(mpaddsqrtrc_loop):
+	move.l -(%a1),%d1
+dnl square %d1 into %d0 and %d1
+	mulu.l %d1,%d0:%d1
+	add.l %d3,%d1
+	addx.l %d4,%d0
+	add.l -(%a0),%d1
+	addx.l %d4,%d0
+	move.l %d1,(%a0)
+	clr.l %d3
+	add.l -(%a0),%d0
+	addx.l %d4,%d3
+	move.l %d0,0(%a0)
+	dbf %d5,LOCAL(mpaddsqrtrc_loop)
+
+	movem.l (%sp)+,%d3-%d5
+	rts
+C_FUNCTION_END(mpaddsqrtrc)
diff --git a/beecrypt/gas/mp32opt.powerpc.S b/beecrypt/gas/mpopt.ppc.m4
index b70f07903..3406f3518 100644
--- a/beecrypt/gas/mp32opt.powerpc.S
+++ b/beecrypt/gas/mpopt.ppc.m4
@@ -1,39 +1,28 @@
-/*
- * mp32opt.powerpc.S
- *
- * Assembler optimized multiprecision integer routines for PowerPC
- *
- * Compile target is GNU Assembler
- *
- * Copyright (c) 2000, 2001 Virtual Unlimited B.V.
- *
- * Author: Bob Deblier <bob@virtualunlimited.com>
- *
- * This library is free software; you can redistribute it and/or
- * modify it under the terms of the GNU Lesser General Public
- * License as published by the Free Software Foundation; either
- * version 2.1 of the License, or (at your option) any later version.
- *
- * This library 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
- * Lesser General Public License for more details.
- *
- * You should have received a copy of the GNU Lesser General Public
- * License along with this library; if not, write to the Free Software
- * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
- *
- */
-
-#include "beecrypt.gas.h"
-
-	.file "mp32opt.powerpc.S"
-
-	.text
+dnl  mpopt.ppc.m4
+dnl
+dnl  Copyright (c) 2003 Bob Deblier
+dnl 
+dnl  Author: Bob Deblier <bob.deblier@pandora.be>
+dnl 
+dnl  This library is free software; you can redistribute it and/or
+dnl  modify it under the terms of the GNU Lesser General Public
+dnl  License as published by the Free Software Foundation; either
+dnl  version 2.1 of the License, or (at your option) any later version.
+dnl 
+dnl  This library 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 GNU
+dnl  Lesser General Public License for more details.
+dnl 
+dnl  You shoulwz have received a copy of the GNU Lesser General Public
+dnl  License along with this library; if not, write to the Free Software
+dnl  Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
+ 
+include(config.m4)
+include(ASM_SRCDIR/ppc.m4)
 
 
 C_FUNCTION_BEGIN(mpaddw)
-LABEL(mpaddw)
 	mtctr r3
 	slwi r0,r3,2
 	add r4,r4,r0
@@ -48,14 +37,12 @@ LOCAL(mpaddw_loop):
 	stw r6,0(r4)
 	bdnz LOCAL(mpaddw_loop)
 LOCAL(mpaddw_skip):
-	/* return the carry */
 	addze r3,r0
 	blr
-C_FUNCTION_END(mpaddw, LOCAL(mpaddw_size))
+C_FUNCTION_END(mpaddw)
 
 
 C_FUNCTION_BEGIN(mpsubw)
-LABEL(mpsubw)
 	mtctr r3
 	slwi r0,r3,2
 	add r4,r4,r0
@@ -70,15 +57,13 @@ LOCAL(mpsubw_loop):
 	stwu r6, -4(r4)
 	bdnz LOCAL(mpsubw_loop)
 LOCAL(mpsubw_skip):
-	/* return the carry */
 	subfe r3,r0,r0
 	neg r3,r3
 	blr
-C_FUNCTION_END(mpsubw, LOCAL(mpsubw_size))
+C_FUNCTION_END(mpsubw)
 
 
 C_FUNCTION_BEGIN(mpadd)
-LABEL(mpadd)
 	mtctr r3
 	slwi r0,r3,2
 	add r4,r4,r0
@@ -96,14 +81,12 @@ LOCAL(mpadd_loop):
 	stwu r6,-4(r4)
 	bdnz LOCAL(mpadd_loop)
 LOCAL(mpadd_skip):
-	/* return the carry */
 	addze r3,r0
 	blr
-C_FUNCTION_END(mpadd, LOCAL(mpadd_size))
+C_FUNCTION_END(mpadd)
 
 
 C_FUNCTION_BEGIN(mpsub)
-LABEL(mpsub)
 	mtctr r3
 	slwi r0,r3,2
 	add r4,r4,r0
@@ -121,15 +104,13 @@ LOCAL(mpsub_loop):
 	stwu r6,-4(r4)
 	bdnz LOCAL(mpsub_loop)
 LOCAL(mpsub_skip):
-	/* return the carry */
 	subfe r3,r0,r0
 	neg r3,r3
 	blr
-C_FUNCTION_END(mpsub, LOCAL(mpsub_size))
+C_FUNCTION_END(mpsub)
 
 
 C_FUNCTION_BEGIN(mpmultwo)
-LABEL(mpmultwo)
 	mtctr r3
 	slwi r0,r3,2
 	add r4,r4,r0
@@ -144,14 +125,12 @@ LOCAL(mpmultwo_loop):
 	stwu r6,-4(r4)
 	bdnz LOCAL(mpmultwo_loop)
 LOCAL(mpmultwo_skip):
-	/* return the carry */
 	addze r3,r0
 	blr
-C_FUNCTION_END(mpmultwo, LOCAL(mpmultwo_size))
+C_FUNCTION_END(mpmultwo)
 
 
 C_FUNCTION_BEGIN(mpsetmul)
-LABEL(mpsetmul)
 	mtctr r3
 	slwi r0,r3,2
 	add r4,r4,r0
@@ -166,11 +145,10 @@ LOCAL(mpsetmul_loop):
 	stwu r8,-4(r4)
 	bdnz LOCAL(mpsetmul_loop)
 	blr
-C_FUNCTION_END(mpsetmul, LOCAL(mpsetmul_size))
+C_FUNCTION_END(mpsetmul)
 
 
 C_FUNCTION_BEGIN(mpaddmul)
-LABEL(mpaddmul)
 	mtctr r3
 	slwi r0,r3,2
 	add r4,r4,r0
@@ -188,11 +166,10 @@ LOCAL(mpaddmul_loop):
 	stw r9,0(r4)
 	bdnz LOCAL(mpaddmul_loop)
 	blr
-C_FUNCTION_END(mpaddmul, LOCAL(mpaddmul_size))
+C_FUNCTION_END(mpaddmul)
 
 
 C_FUNCTION_BEGIN(mpaddsqrtrc)
-LABEL(mpaddsqrtrc)
 	mtctr r3
 	slwi r0,r3,2
 	add r4,r4,r0
@@ -215,4 +192,4 @@ LOCAL(mpaddsqrtrc_loop):
 	stwu r6,-8(r4)
 	bdnz LOCAL(mpaddsqrtrc_loop)
 	blr
-C_FUNCTION_END(mpaddsqrtrc, LOCAL(mpaddsqrtrc_size))
+C_FUNCTION_END(mpaddsqrtrc)
diff --git a/beecrypt/gas/mpopt.ppc64.m4 b/beecrypt/gas/mpopt.ppc64.m4
new file mode 100644
index 000000000..8fdbdb0ae
--- /dev/null
+++ b/beecrypt/gas/mpopt.ppc64.m4
@@ -0,0 +1,195 @@
+dnl  mpopt.ppc64.m4
+dnl
+dnl  Copyright (c) 2003 Bob Deblier
+dnl 
+dnl  Author: Bob Deblier <bob.deblier@pandora.be>
+dnl 
+dnl  This library is free software; you can redistribute it and/or
+dnl  modify it under the terms of the GNU Lesser General Public
+dnl  License as published by the Free Software Foundation; either
+dnl  version 2.1 of the License, or (at your option) any later version.
+dnl 
+dnl  This library 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 GNU
+dnl  Lesser General Public License for more details.
+dnl 
+dnl  You should have received a copy of the GNU Lesser General Public
+dnl  License along with this library; if not, write to the Free Software
+dnl  Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
+ 
+include(config.m4)
+include(ASM_SRCDIR/ppc64.m4)
+
+
+C_FUNCTION_BEGIN(mpaddw)
+	mtctr r3
+	sldi r0,r3,3
+	add r4,r4,r0
+	li r0,0
+	ldu r6,-8(r4)
+	addc r6,r6,r5
+	std r6,0(r4)
+	bdz LOCAL(mpaddw_skip)
+LOCAL(mpaddw_loop):
+	ldu r6,-8(r4)
+	adde r6,r0,r6
+	std r6,0(r4)
+	bdnz LOCAL(mpaddw_loop)
+LOCAL(mpaddw_skip):
+	addze r3,r0
+	blr
+C_FUNCTION_END(mpaddw)
+
+
+C_FUNCTION_BEGIN(mpsubw)
+	mtctr r3
+	sldi r0,r3,3
+	add r4,r4,r0
+	li r0,0
+	ld r6,-8(r4)
+	subfc r6,r5,r6
+	stdu r6,-8(r4)
+	bdz LOCAL(mpsubw_skip)
+LOCAL(mpsubw_loop):
+	ld r6,-8(r4)
+	subfe r6,r0,r6
+	stdu r6, -8(r4)
+	bdnz LOCAL(mpsubw_loop)
+LOCAL(mpsubw_skip):
+	subfe r3,r0,r0
+	neg r3,r3
+	blr
+C_FUNCTION_END(mpsubw)
+
+
+C_FUNCTION_BEGIN(mpadd)
+	mtctr r3
+	sldi r0,r3,3
+	add r4,r4,r0
+	add r5,r5,r0
+	li r0,0
+	ld r6,-8(r4)
+	ldu r7,-8(r5)
+	addc r6,r7,r6
+	stdu r6,-8(r4)
+	bdz LOCAL(mpadd_skip)
+LOCAL(mpadd_loop):
+	ld r6,-8(r4)
+	ldu r7,-8(r5)
+	adde r6,r7,r6
+	stdu r6,-8(r4)
+	bdnz LOCAL(mpadd_loop)
+LOCAL(mpadd_skip):
+	addze r3,r0
+	blr
+C_FUNCTION_END(mpadd)
+
+
+C_FUNCTION_BEGIN(mpsub)
+	mtctr r3
+	sldi r0,r3,3
+	add r4,r4,r0
+	add r5,r5,r0
+	li r0,0
+	ld r6,-8(r4)
+	ldu r7,-8(r5)
+	subfc r6,r7,r6
+	stdu r6,-8(r4)
+	bdz LOCAL(mpsub_skip)
+LOCAL(mpsub_loop):
+	ld r6,-8(r4)
+	ldu r7,-8(r5)
+	subfe r6,r7,r6
+	stdu r6,-8(r4)
+	bdnz LOCAL(mpsub_loop)
+LOCAL(mpsub_skip):
+	subfe r3,r0,r0
+	neg r3,r3
+	blr
+C_FUNCTION_END(mpsub)
+
+
+C_FUNCTION_BEGIN(mpmultwo)
+	mtctr r3
+	sldi r0,r3,3
+	add r4,r4,r0
+	li r0,0
+	ld r6,-8(r4)
+	addc r6,r6,r6
+	stdu r6,-8(r4)
+	bdz LOCAL(mpmultwo_skip)
+LOCAL(mpmultwo_loop):
+	ld r6,-8(r4)
+	adde r6,r6,r6
+	stdu r6,-8(r4)
+	bdnz LOCAL(mpmultwo_loop)
+LOCAL(mpmultwo_skip):
+	addze r3,r0
+	blr
+C_FUNCTION_END(mpmultwo)
+
+
+C_FUNCTION_BEGIN(mpsetmul)
+	mtctr r3
+	sldi r0,r3,3
+	add r4,r4,r0
+	add r5,r5,r0
+	li r3,0
+LOCAL(mpsetmul_loop):
+	ldu r7,-8(r5)
+	mulld r8,r7,r6
+	addc r8,r8,r3
+	mulhdu r9,r7,r6
+	addze r3,r9
+	stdu r8,-8(r4)
+	bdnz LOCAL(mpsetmul_loop)
+	blr
+C_FUNCTION_END(mpsetmul)
+
+
+C_FUNCTION_BEGIN(mpaddmul)
+	mtctr r3
+	sldi r0,r3,3
+	add r4,r4,r0
+	add r5,r5,r0
+	li r3,0
+LOCAL(mpaddmul_loop):
+	ldu r8,-8(r5)
+	ldu r7,-8(r4)
+	mulld r9,r8,r6
+	addc r9,r9,r3
+	mulhdu r10,r8,r6
+	addze r3,r10
+	addc r9,r9,r7
+	addze r3,r3
+	std r9,0(r4)
+	bdnz LOCAL(mpaddmul_loop)
+	blr
+C_FUNCTION_END(mpaddmul)
+
+
+C_FUNCTION_BEGIN(mpaddsqrtrc)
+	mtctr r3
+	sldi r0,r3,3
+	add r4,r4,r0
+	add r5,r5,r0
+	add r4,r4,r0
+	li r3,0
+LOCAL(mpaddsqrtrc_loop):
+	ldu r0,-8(r5)
+	ld r6,-16(r4)
+	ld r7,-8(r4)
+	mulld r9,r0,r0
+	addc r9,r9,r3
+	mulhdu r8,r0,r0
+	addze r8,r8
+	li r3,0
+	addc r7,r7,r9
+	adde r6,r6,r8
+	addze r3,r3
+	std r7,-8(r4)
+	stdu r6,-16(r4)
+	bdnz LOCAL(mpaddsqrtrc_loop)
+	blr
+C_FUNCTION_END(mpaddsqrtrc)
diff --git a/beecrypt/gas/mpopt.sparcv8.m4 b/beecrypt/gas/mpopt.sparcv8.m4
new file mode 100644
index 000000000..f21b35614
--- /dev/null
+++ b/beecrypt/gas/mpopt.sparcv8.m4
@@ -0,0 +1,90 @@
+dnl  mpopt.sparcv8.m4
+dnl
+dnl  Copyright (c) 2003 Bob Deblier
+dnl 
+dnl  Author: Bob Deblier <bob.deblier@pandora.be>
+dnl 
+dnl  This library is free software; you can redistribute it and/or
+dnl  modify it under the terms of the GNU Lesser General Public
+dnl  License as published by the Free Software Foundation; either
+dnl  version 2.1 of the License, or (at your option) any later version.
+dnl 
+dnl  This library 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 GNU
+dnl  Lesser General Public License for more details.
+dnl 
+dnl  You should have received a copy of the GNU Lesser General Public
+dnl  License along with this library; if not, write to the Free Software
+dnl  Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
+
+include(config.m4)
+include(ASM_SRCDIR/sparc.m4)
+
+
+C_FUNCTION_BEGIN(mpsetmul)
+	sll %o0,2,%g1
+	dec 4,%o2
+	clr %o0
+LOCAL(mpsetmul_loop):
+	ld [%o2+%g1],%g2
+	umul %o3,%g2,%g2
+	rd %y,%g3
+	addcc %o0,%g2,%g2
+	addx %g0,%g3,%o0
+	deccc 4,%g1
+	bnz LOCAL(mpsetmul_loop)
+	st %g2,[%o1+%g1]
+	retl
+	nop
+C_FUNCTION_END(mpsetmul)
+
+
+C_FUNCTION_BEGIN(mpaddmul)
+	sll %o0,2,%g1
+	mov %o1,%o4
+	dec 4,%o1
+	dec 4,%o2
+	clr %o0
+LOCAL(mpaddmul_loop):
+	ld [%o2+%g1],%g2
+	ld [%o1+%g1],%g3
+	umul %o3,%g2,%g2
+	rd %y,%g4
+	addcc %o0,%g2,%g2
+	addx %g0,%g4,%g4
+	addcc %g2,%g3,%g2
+	addx %g0,%g4,%o0
+	deccc 4,%g1
+	bnz LOCAL(mpaddmul_loop)
+	st %g2,[%o4+%g1]
+	retl
+	nop
+C_FUNCTION_END(mpaddmul)
+
+
+C_FUNCTION_BEGIN(mpaddsqrtrc)
+        sll %o0,2,%g1
+        add %o1,%g1,%o1
+		dec 4,%o2
+        add %o1,%g1,%o1
+		dec 8,%o1
+        clr %o0
+LOCAL(mpaddsqrtrc_loop):
+        ld [%o2+%g1],%g2
+        ldd [%o1],%o4
+        umul %g2,%g2,%g3
+        rd %y,%g2
+        addcc %o5,%g3,%o5
+        addxcc %o4,%g2,%o4
+        addx %g0,%g0,%o3
+        addcc %o5,%o0,%o5
+        addxcc %o4,%g0,%o4
+        addx %o3,%g0,%o0
+        std %o4,[%o1]
+        deccc 4,%g1
+        bnz LOCAL(mpaddsqrtrc_loop)
+        sub %o1,8,%o1
+        retl
+        nop
+C_FUNCTION_END(mpaddsqrtrc)
diff --git a/beecrypt/gas/mp32opt.sparcv9.S b/beecrypt/gas/mpopt.sparcv8plus.m4
index 582a68641..f021cfa6c 100644
--- a/beecrypt/gas/mp32opt.sparcv9.S
+++ b/beecrypt/gas/mpopt.sparcv8plus.m4
@@ -1,41 +1,28 @@
-/*
- * mp32opt.sparcv9.S
- *
- * Assembler optimized multiprecision integer routines for UltraSparc (64 bits instructions, will run on 32 bit OS)
- *
- * Compile target is GNU Assembler, Sun Solaris Assembler
- *
- * Copyright (c) 1998, 1999, 2000, 2001 Virtual Unlimited B.V.
- *
- * Author: Bob Deblier <bob@virtualunlimited.com>
- *
- * This library is free software; you can redistribute it and/or
- * modify it under the terms of the GNU Lesser General Public
- * License as published by the Free Software Foundation; either
- * version 2.1 of the License, or (at your option) any later version.
- *
- * This library 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
- * Lesser General Public License for more details.
- *
- * You should have received a copy of the GNU Lesser General Public
- * License along with this library; if not, write to the Free Software
- * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
- *
- */
-
-#include "beecrypt.gas.h"
-
-	.file "mp32opt.sparcv9.S"
-
-	.text
+dnl  mpopt.sparcv8plus.m4
+dnl
+dnl  Copyright (c) 2003 Bob Deblier
+dnl 
+dnl  Author: Bob Deblier <bob.deblier@pandora.be>
+dnl 
+dnl  This library is free software; you can redistribute it and/or
+dnl  modify it under the terms of the GNU Lesser General Public
+dnl  License as published by the Free Software Foundation; either
+dnl  version 2.1 of the License, or (at your option) any later version.
+dnl 
+dnl  This library 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 GNU
+dnl  Lesser General Public License for more details.
+dnl 
+dnl  You should have received a copy of the GNU Lesser General Public
+dnl  License along with this library; if not, write to the Free Software
+dnl  Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
+
+include(config.m4)
+include(ASM_SRCDIR/sparc.m4)
 
-C_FUNCTION_BEGIN(mpaddw)
-LABEL(mpaddw)
-
-	.register %g2,#scratch
 
+C_FUNCTION_BEGIN(mpaddw)
 	sll %o0,2,%g1
 	dec 4,%g1
 	clr %o0
@@ -53,14 +40,10 @@ LOCAL(mpaddw_loop):
 LOCAL(mpaddw_skip):
 	retl
 	movcs %icc,1,%o0
-C_FUNCTION_END(mpaddw, LOCAL(mpaddw_size))
+C_FUNCTION_END(mpaddw)
 
 
 C_FUNCTION_BEGIN(mpsubw)
-LABEL(mpsubw)
-
-	.register %g2,#scratch
-
 	sll %o0,2,%g1
 	dec 4,%g1
 	clr %o0
@@ -78,15 +61,10 @@ LOCAL(mpsubw_loop):
 LOCAL(mpsubw_skip):
 	retl
 	movcs %icc,1,%o0
-C_FUNCTION_END(mpsubw, LOCAL(mpsubw_size))
+C_FUNCTION_END(mpsubw)
 
 
 C_FUNCTION_BEGIN(mpadd)
-LABEL(mpadd)
-
-	.register %g2,#scratch
-	.register %g3,#scratch
-
 	sll %o0,2,%g1
 	dec 4,%g1
 	addcc %g0,%g0,%o0
@@ -99,15 +77,10 @@ LOCAL(mpadd_loop):
 	dec 4,%g1
 	retl
 	movcs %icc,1,%o0
-C_FUNCTION_END(mpadd, LOCAL(mpadd_size))
+C_FUNCTION_END(mpadd)
 
 
 C_FUNCTION_BEGIN(mpsub)
-LABEL(mpsub)
-
-	.register %g2,#scratch
-	.register %g3,#scratch
-
 	sll %o0,2,%g1
 	dec 4,%g1
 	addcc %g0,%g0,%o0
@@ -120,15 +93,10 @@ LOCAL(mpsub_loop):
 	dec 4,%g1
 	retl
 	movcs %icc,1,%o0
-C_FUNCTION_END(mpsub, LOCAL(mpsub_size))
+C_FUNCTION_END(mpsub)
 
 
 C_FUNCTION_BEGIN(mpmultwo)
-LABEL(mpmultwo)
-
-	.register %g2,#scratch
-	.register %g3,#scratch
-
 	sll %o0,2,%g1
 	dec 4,%g1
 	addcc %g0,%g0,%o0
@@ -140,15 +108,10 @@ LOCAL(mpmultwo_loop):
 	dec 4,%g1
 	retl
 	movcs %icc,1,%o0
-C_FUNCTION_END(mpmultwo, LOCAL(mpmultwo_size))
+C_FUNCTION_END(mpmultwo)
 
 
 C_FUNCTION_BEGIN(mpsetmul)
-LABEL(mpsetmul)
-
-	.register %g2,#scratch
-	.register %g3,#scratch
-
 	sll %o0,2,%g1
 	dec 4,%g1
 	clr %o0
@@ -162,15 +125,10 @@ LOCAL(mpsetmul_loop):
 	dec 4,%g1
 	retl
 	srlx %o0,32,%o0
-C_FUNCTION_END(mpsetmul, LOCAL(mpsetmul_size))
+C_FUNCTION_END(mpsetmul)
 
 
 C_FUNCTION_BEGIN(mpaddmul)
-LABEL(mpaddmul)
-
-	.register %g2,#scratch
-	.register %g3,#scratch
-
 	sll %o0,2,%g1
 	dec 4,%g1
 	clr %o0
@@ -186,22 +144,16 @@ LOCAL(mpaddmul_loop):
 	dec 4,%g1
 	retl
 	srlx %o0,32,%o0
-C_FUNCTION_END(mpaddmul, LOCAL(mpaddmul_size))
+C_FUNCTION_END(mpaddmul)
 
 
 C_FUNCTION_BEGIN(mpaddsqrtrc)
-LABEL(mpaddsqrtrc)
-
-	.register %g2,#scratch
-	.register %g3,#scratch
-
 	sll %o0,2,%g1
 	dec 4,%g1
 	add %o1,%g1,%o1
 	add %o1,%g1,%o1
 	clr %o0
 LOCAL(mpaddsqrtrc_loop):
-	/* load from o1 into g4 as xuint; simulate xuint carry by doing an xuint comparison; carry if result smaller than initial value */
 	lduw [%o2+%g1],%g2
 	ldx [%o1],%g4
 	mulx %g2,%g2,%g2
@@ -216,4 +168,4 @@ LOCAL(mpaddsqrtrc_loop):
 	dec 4,%g1
 	retl
 	nop
-C_FUNCTION_END(mpaddsqrtrc, LOCAL(mpaddsqrtrc_size))
+C_FUNCTION_END(mpaddsqrtrc)
diff --git a/beecrypt/gas/mp32opt.i386.S b/beecrypt/gas/mpopt.x86.m4
index 967faa82b..e1097999d 100644
--- a/beecrypt/gas/mp32opt.i386.S
+++ b/beecrypt/gas/mpopt.x86.m4
@@ -1,39 +1,28 @@
-/*
- * mp32opt.i386.S
- *
- * Assembler optimized multiprecision integer routines for Intel 386 and higher
- *
- * Compile target is GNU Assembler
- *
- * Copyright (c) 1998, 1999, 2000, 2001 Virtual Unlimited B.V.
- *
- * Author: Bob Deblier <bob@virtualunlimited.com>
- *
- * This library is free software; you can redistribute it and/or
- * modify it under the terms of the GNU Lesser General Public
- * License as published by the Free Software Foundation; either
- * version 2.1 of the License, or (at your option) any later version.
- *
- * This library 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
- * Lesser General Public License for more details.
- *
- * You should have received a copy of the GNU Lesser General Public
- * License along with this library; if not, write to the Free Software
- * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
- *
- */
-
-#include "beecrypt.gas.h"
-
-	.file "mp32opt.i386.S"
-
-	.text
+dnl  mpopt.x86.m4
+dnl
+dnl  Copyright (c) 2003 Bob Deblier
+dnl 
+dnl  Author: Bob Deblier <bob.deblier@pandora.be>
+dnl 
+dnl  This library is free software; you can redistribute it and/or
+dnl  modify it under the terms of the GNU Lesser General Public
+dnl  License as published by the Free Software Foundation; either
+dnl  version 2.1 of the License, or (at your option) any later version.
+dnl 
+dnl  This library 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 GNU
+dnl  Lesser General Public License for more details.
+dnl 
+dnl  You should have received a copy of the GNU Lesser General Public
+dnl  License along with this library; if not, write to the Free Software
+dnl  Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
+ 
+include(config.m4)
+include(ASM_SRCDIR/x86.m4)
 
 
 C_FUNCTION_BEGIN(mpzero)
-LABEL(mpzero)
 	pushl %edi
 
 	movl 8(%esp),%ecx
@@ -44,11 +33,10 @@ LABEL(mpzero)
 
 	popl %edi
 	ret
-C_FUNCTION_END(mpzero, LOCAL(mpzero_size))
+C_FUNCTION_END(mpzero)
 
 
 C_FUNCTION_BEGIN(mpfill)
-LABEL(mpfill)
 	pushl %edi
 
 	movl 8(%esp),%ecx
@@ -59,32 +47,29 @@ LABEL(mpfill)
 
 	popl %edi
 	ret
-C_FUNCTION_END(mpfill, LOCAL(mpfill_size))
+C_FUNCTION_END(mpfill)
 
 
 C_FUNCTION_BEGIN(mpeven)
-LABEL(mpeven)
 	movl 4(%esp),%ecx
 	movl 8(%esp),%eax
 	movl -4(%eax,%ecx,4),%eax
 	notl %eax
-	andl $1,%eax
+	andl `$'1,%eax
 	ret
-C_FUNCTION_END(mpeven, LOCAL(mpeven_size))
+C_FUNCTION_END(mpeven)
 
 
 C_FUNCTION_BEGIN(mpodd)
-LABEL(mpodd)
 	movl 4(%esp),%ecx
 	movl 8(%esp),%eax
 	movl -4(%eax,%ecx,4),%eax
-	andl $1,%eax
+	andl `$'1,%eax
 	ret
-C_FUNCTION_END(mpodd, LOCAL(mpodd_size))
+C_FUNCTION_END(mpodd)
 
 
 C_FUNCTION_BEGIN(mpaddw)
-LABEL(mpaddw)
 	pushl %edi
 
 	movl 8(%esp),%ecx
@@ -110,11 +95,10 @@ LOCAL(mpaddw_skip):
 
 	popl %edi
 	ret
-C_FUNCTION_END(mpaddw, LOCAL(mpaddw_size))
+C_FUNCTION_END(mpaddw)
 
 
 C_FUNCTION_BEGIN(mpsubw)
-LABEL(mpsubw)
 	pushl %edi
 
 	movl 8(%esp),%ecx
@@ -139,11 +123,10 @@ LOCAL(mpsubw_skip):
 	negl %eax
 	popl %edi
 	ret
-C_FUNCTION_END(mpsubw, LOCAL(mpsubw_size))
+C_FUNCTION_END(mpsubw)
 
 
 C_FUNCTION_BEGIN(mpadd)
-LABEL(mpadd)
 	pushl %edi
 	pushl %esi
 
@@ -157,7 +140,9 @@ LABEL(mpadd)
 	.align 4
 LOCAL(mpadd_loop):
 	movl (%esi,%ecx,4),%eax
-	adcl %eax,(%edi,%ecx,4)
+	movl (%edi,%ecx,4),%edx
+	adcl %eax,%edx
+	movl %edx,(%edi,%ecx,4)
 	decl %ecx
 	jns LOCAL(mpadd_loop)
 
@@ -167,11 +152,10 @@ LOCAL(mpadd_loop):
 	popl %esi
 	popl %edi
 	ret
-C_FUNCTION_END(mpadd, LOCAL(mpadd_size))
+C_FUNCTION_END(mpadd)
 
 
 C_FUNCTION_BEGIN(mpsub)
-LABEL(mpsub)
 	pushl %edi
 	pushl %esi
 
@@ -185,7 +169,9 @@ LABEL(mpsub)
 	.align 4
 LOCAL(mpsub_loop):
 	movl (%esi,%ecx,4),%eax
-	sbbl %eax,(%edi,%ecx,4)
+	movl (%edi,%ecx,4),%edx
+	sbbl %eax,%edx
+	movl %edx,(%edi,%ecx,4)
 	decl %ecx
 	jns LOCAL(mpsub_loop)
 
@@ -194,11 +180,10 @@ LOCAL(mpsub_loop):
 	popl %esi
 	popl %edi
 	ret
-C_FUNCTION_END(mpsub, LOCAL(mpsub_size))
+C_FUNCTION_END(mpsub)
 
 
 C_FUNCTION_BEGIN(mpdivtwo)
-LABEL(mpdivtwo)
 	pushl %edi
 
 	movl 8(%esp),%ecx
@@ -206,32 +191,33 @@ LABEL(mpdivtwo)
 
 	leal (%edi,%ecx,4),%edi
 	negl %ecx
-	clc
+	xorl %eax,%eax
 
 	.align 4
 LOCAL(mpdivtwo_loop):
-	rcrl $1,(%edi,%ecx,4)
+	rcrl `$'1,(%edi,%ecx,4)
 	inc %ecx
 	jnz LOCAL(mpdivtwo_loop)
 
 	popl %edi
 	ret
-C_FUNCTION_END(mpdivtwo, LOCAL(mpdivtwo_size))
+C_FUNCTION_END(mpdivtwo)
 
 
 C_FUNCTION_BEGIN(mpmultwo)
-LABEL(mpmultwo)
 	pushl %edi
 
 	movl 8(%esp),%ecx
 	movl 12(%esp),%edi
 
-	clc
+	xorl %edx,%edx
 	decl %ecx
 
 	.align 4
 LOCAL(mpmultwo_loop):
-	rcll $1,(%edi,%ecx,4)
+	movl (%edi,%ecx,4),%eax
+	adcl %eax,%eax
+	movl %eax,(%edi,%ecx,4)
 	decl %ecx 
 	jns LOCAL(mpmultwo_loop)
 
@@ -240,13 +226,34 @@ LOCAL(mpmultwo_loop):
 
 	popl %edi
 	ret
-C_FUNCTION_END(mpmultwo, LOCAL(mpmultwo_size))
+C_FUNCTION_END(mpmultwo)
 
 
 C_FUNCTION_BEGIN(mpsetmul)
-LABEL(mpsetmul)
 	pushl %edi
 	pushl %esi
+ifdef(`USE_SSE2',`
+	movl 12(%esp),%ecx
+	movl 16(%esp),%edi
+	movl 20(%esp),%esi
+	movd 24(%esp),%mm1
+
+	pxor %mm0,%mm0
+	decl %ecx
+
+	.align 4
+LOCAL(mpsetmul_loop):
+	movd (%esi,%ecx,4),%mm2
+	pmuludq %mm1,%mm2
+	paddq %mm2,%mm0
+	movd %mm0,(%edi,%ecx,4)
+	decl %ecx
+	psrlq `$'32,%mm0
+	jns LOCAL(mpsetmul_loop)
+
+	movd %mm0,%eax
+	emms
+',`
 	pushl %ebx
 	pushl %ebp
 
@@ -264,7 +271,7 @@ LOCAL(mpsetmul_loop):
 	movl (%esi,%ecx,4),%eax
 	mull %ebp
 	addl %ebx,%eax
-	adcl $0,%edx
+	adcl `$'0,%edx
 	movl %eax,(%edi,%ecx,4)
 	decl %ecx
 	jns LOCAL(mpsetmul_loop)
@@ -273,16 +280,40 @@ LOCAL(mpsetmul_loop):
 
 	popl %ebp
 	popl %ebx
+')
 	popl %esi
 	popl %edi
 	ret
-C_FUNCTION_END(mpsetmul, LOCAL(mpsetmul_size))
+C_FUNCTION_END(mpsetmul)
 
 
 C_FUNCTION_BEGIN(mpaddmul)
-LABEL(mpaddmul)
 	pushl %edi
 	pushl %esi
+ifdef(`USE_SSE2',`
+	movl 12(%esp),%ecx
+	movl 16(%esp),%edi
+	movl 20(%esp),%esi
+	movd 24(%esp),%mm1
+
+	pxor %mm0,%mm0
+	decl %ecx
+
+	.align 4
+LOCAL(mpaddmul_loop):
+	movd (%esi,%ecx,4),%mm2
+	movd (%edi,%ecx,4),%mm3
+	pmuludq %mm1,%mm2
+	paddq %mm2,%mm3
+	paddq %mm3,%mm0
+	movd %mm0,(%edi,%ecx,4)
+	decl %ecx
+	psrlq $32,%mm0
+	jns LOCAL(mpaddmul_loop)
+
+	movd %mm0,%eax
+	emms
+',`
 	pushl %ebx
 	pushl %ebp
 
@@ -311,16 +342,43 @@ LOCAL(mpaddmul_loop):
 
 	popl %ebp
 	popl %ebx
+')
 	popl %esi
 	popl %edi
 	ret
-C_FUNCTION_END(mpaddmul, LOCAL(mpaddmul_size))
+C_FUNCTION_END(mpaddmul)
 
 
 C_FUNCTION_BEGIN(mpaddsqrtrc)
-LABEL(mpaddsqrtrc)
 	pushl %edi
 	pushl %esi
+ifdef(`USE_SSE2',`
+	movl 12(%esp),%ecx
+	movl 16(%esp),%edi
+	movl 20(%esp),%esi
+
+	pxor %mm0,%mm0
+	decl %ecx
+
+	.align 4
+LOCAL(mpaddsqrtrc_loop):
+	movd (%esi,%ecx,4),%mm2
+	pmuludq %mm2,%mm2
+	movd 4(%edi,%ecx,8),%mm3
+	paddq %mm2,%mm3
+	movd 0(%edi,%ecx,8),%mm4
+	paddq %mm3,%mm0
+	movd %mm0,4(%edi,%ecx,8)
+	psrlq $32,%mm0
+	paddq %mm4,%mm0
+	movd %mm0,0(%edi,%ecx,8)
+	decl %ecx
+	psrlq $32,%mm0
+	jns LOCAL(mpaddsqrtrc_loop)
+
+	movd %mm0,%eax
+	emms
+',`
 	pushl %ebx
 
 	movl 16(%esp),%ecx
@@ -346,7 +404,8 @@ LOCAL(mpaddsqrtrc_loop):
 	movl %ebx,%eax
 
 	popl %ebx
+')
 	popl %esi
 	popl %edi
 	ret
-C_FUNCTION_END(mpaddsqrtrc, LOCAL(mpaddsqrtrc_size))
+C_FUNCTION_END(mpaddsqrtrc)
diff --git a/beecrypt/gas/ppc.m4 b/beecrypt/gas/ppc.m4
new file mode 100644
index 000000000..f8e9865db
--- /dev/null
+++ b/beecrypt/gas/ppc.m4
@@ -0,0 +1,85 @@
+dnl  ppc.m4
+dnl
+dnl  Copyright (c) 2003 Bob Deblier
+dnl 
+dnl  Author: Bob Deblier <bob.deblier@pandora.be>
+dnl 
+dnl  This library is free software; you can redistribute it and/or
+dnl  modify it under the terms of the GNU Lesser General Public
+dnl  License as published by the Free Software Foundation; either
+dnl  version 2.1 of the License, or (at your option) any later version.
+dnl 
+dnl  This library 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 GNU
+dnl  Lesser General Public License for more details.
+dnl 
+dnl  You should have received a copy of the GNU Lesser General Public
+dnl  License along with this library; if not, write to the Free Software
+dnl  Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
+
+ifelse(substr(ASM_OS,0,3),aix,`
+undefine(`C_FUNCTION_BEGIN')
+define(C_FUNCTION_BEGIN,`
+	.toc
+	.globl	$1[DS]
+	.csect	$1[DS]
+	.long	.$1[PR], TOC[tc0], 0
+	.toc
+	.globl	.$1[PR]
+	.csect	.$1[PR]
+')
+undefine(`C_FUNCTION_END')
+define(C_FUNCTION_END,`
+	.tbtag 0x0,0xc,0x0,0x0,0x0,0x0,0x0,0x0
+')
+define(LOAD_ADDRESS,`
+	lwz $2,L$1(r2)
+')
+define(EXTERNAL_VARIABLE,`
+	.toc
+L$1:
+	.tc $1[TC],$1[RW]
+')
+
+	.machine	"ppc"
+
+	.set r0,0
+	.set r1,1
+	.set r2,2
+	.set r3,3
+	.set r4,4
+	.set r5,5
+	.set r6,6
+	.set r7,7
+	.set r8,8
+	.set r9,9
+	.set r10,10
+	.set r11,11
+	.set r12,12
+	.set r13,13
+	.set r14,14
+	.set r15,15
+	.set r16,16
+	.set r17,17
+	.set r18,18
+	.set r19,19
+	.set r20,20
+	.set r21,21
+	.set r22,22
+	.set r23,23
+	.set r24,24
+	.set r25,25
+	.set r26,26
+	.set r27,27
+	.set r28,28
+	.set r29,29
+	.set r30,30
+	.set r31,31
+',`
+define(LOAD_ADDRESS,`
+	lis $2,$1@ha
+	la $2,$1@l($2)
+')
+define(EXTERNAL_VARIABLE)
+')
diff --git a/beecrypt/gas/ppc64.m4 b/beecrypt/gas/ppc64.m4
new file mode 100644
index 000000000..38bec0e47
--- /dev/null
+++ b/beecrypt/gas/ppc64.m4
@@ -0,0 +1,71 @@
+dnl  ppc64.m4
+dnl
+dnl  Copyright (c) 2003 Bob Deblier
+dnl 
+dnl  Author: Bob Deblier <bob.deblier@pandora.be>
+dnl 
+dnl  This library is free software; you can redistribute it and/or
+dnl  modify it under the terms of the GNU Lesser General Public
+dnl  License as published by the Free Software Foundation; either
+dnl  version 2.1 of the License, or (at your option) any later version.
+dnl 
+dnl  This library 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 GNU
+dnl  Lesser General Public License for more details.
+dnl 
+dnl  You should have received a copy of the GNU Lesser General Public
+dnl  License along with this library; if not, write to the Free Software
+dnl  Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
+
+ifelse(substr(ASM_OS,0,3),aix,`
+undefine(`C_FUNCTION_BEGIN')
+define(C_FUNCTION_BEGIN,`
+	.toc
+	.globl	$1[DS]
+	.csect	$1[DS]
+	.llong	.$1[PR], TOC[tc0], 0
+	.toc
+	.globl	.$1[PR]
+	.csect	.$1[PR]
+')
+undefine(`C_FUNCTION_END')
+define(C_FUNCTION_END,`
+	.tbtag 0x0,0xc,0x0,0x0,0x0,0x0,0x0,0x0
+')
+
+	.machine	"ppc64"
+
+	.set r0,0
+	.set r1,1
+	.set r2,2
+	.set r3,3
+	.set r4,4
+	.set r5,5
+	.set r6,6
+	.set r7,7
+	.set r8,8
+	.set r9,9
+	.set r10,10
+	.set r11,11
+	.set r12,12
+	.set r13,13
+	.set r14,14
+	.set r15,15
+	.set r16,16
+	.set r17,17
+	.set r18,18
+	.set r19,19
+	.set r20,20
+	.set r21,21
+	.set r22,22
+	.set r23,23
+	.set r24,24
+	.set r25,25
+	.set r26,26
+	.set r27,27
+	.set r28,28
+	.set r29,29
+	.set r30,30
+	.set r31,31
+')
diff --git a/beecrypt/gas/sha1opt.i586.S b/beecrypt/gas/sha1opt.i586.S
deleted file mode 100644
index 31e39822b..000000000
--- a/beecrypt/gas/sha1opt.i586.S
+++ /dev/null
@@ -1,292 +0,0 @@
-/*
- * fips180opt.i586.S
- *
- * Assembler optimized SHA-1 routines for Intel Pentium processors
- *
- * Compile target is GNU Assembler
- *
- * Copyright (c) 2000 Virtual Unlimited B.V.
- *
- * Author: Bob Deblier <bob@virtualunlimited.com>
- *
- * This library is free software; you can redistribute it and/or
- * modify it under the terms of the GNU Lesser General Public
- * License as published by the Free Software Foundation; either
- * version 2.1 of the License, or (at your option) any later version.
- *
- * This library 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
- * Lesser General Public License for more details.
- *
- * You should have received a copy of the GNU Lesser General Public
- * License along with this library; if not, write to the Free Software
- * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
- *
- */
-
-#include "beecrypt.gas.h"
-
-	.file "fips180opt.i586.S"
-
-	.text
-
-#define K00	0x5a827999
-#define K20	0x6ed9eba1
-#define K40	0x8f1bbcdc
-#define K60	0xca62c1d6
-
-#define PARAM_H		0
-#define PARAM_DATA	20
-
-	.macro	subround1 b c d e w
-	movl \c,%ecx
-	movl \b,%ebx
-	movl \d,%edx
-	roll $5,%eax
-	xorl %edx,%ecx
-	addl \e,%eax
-	andl %ebx,%ecx
-	addl $K00,%eax
-	rorl $2,%ebx
-	addl \w(%esi,%edi),%eax
-	xorl %edx,%ecx
-	movl %ebx,\b
-	addl %ecx,%eax
-	movl %eax,\e
-	.endm
-
-	.macro subround2 b c d e w
-	movl \c,%ecx
-	movl \b,%ebx
-	roll $5,%eax
-	xorl %ebx,%ecx
-	addl \e,%eax
-	xorl \d,%ecx
-	addl $K20,%eax
-	rorl $2,%ebx
-	addl \w(%esi,%edi),%eax
-	movl %ebx,\b
-	addl %ecx,%eax
-	movl %eax,\e
-	.endm
-
-	.macro subround3 b c d e w
-	movl \c,%ecx
-	roll $5,%eax
-	movl \b,%ebx
-	movl %ecx,%edx
-	addl \e,%eax
-	orl %ebx,%ecx
-	andl %ebx,%edx
-	andl \d,%ecx
-	addl $K40,%eax
-	orl %edx,%ecx
-	addl \w(%esi,%edi),%eax
-	rorl $2,%ebx
-	addl %ecx,%eax
-	movl %ebx,\b
-	movl %eax,\e
-	.endm
-
-	.macro subround4 b c d e w
-	movl \c,%ecx
-	movl \b,%ebx
-	roll $5,%eax
-	xorl %ebx,%ecx
-	addl \e,%eax
-	xorl \d,%ecx
-	addl $K60,%eax
-	rorl $2,%ebx
-	addl \w(%esi,%edi),%eax
-	movl %ebx,\b
-	addl %ecx,%eax
-	movl %eax,\e
-	.endm
-
-C_FUNCTION_BEGIN(sha1Process)
-LABEL(sha1Process)
-	pushl %edi
-	pushl %esi
-	pushl %ebx
-	pushl %ebp
-
-	movl 20(%esp),%esi
-	subl $20,%esp
-	leal PARAM_DATA(%esi),%edi
-	movl %esp,%ebp
-
-	movl $4,%ecx
-LOCAL(0):
-	movl (%esi,%ecx,4),%edx
-	movl %edx,(%ebp,%ecx,4)
-	decl %ecx
-	jns LOCAL(0)
-
-	movl $15,%ecx
-	xorl %eax,%eax
-
-	.p2align 2
-LOCAL(1):
-	movl (%edi,%ecx,4),%edx
-	bswap %edx
-	mov %edx,(%edi,%ecx,4)
-	decl %ecx
-	jns LOCAL(1)
-
-	leal PARAM_DATA(%esi),%edi
-	movl $16,%ecx
-
-	.p2align 2
-LOCAL(2):
-	movl 52(%edi),%eax
-	movl 56(%edi),%ebx
-	xorl 32(%edi),%eax
-	xorl 36(%edi),%ebx
-	xorl 8(%edi),%eax
-	xorl 12(%edi),%ebx
-	xorl (%edi),%eax
-	xorl 4(%edi),%ebx
-	roll $1,%eax
-	roll $1,%ebx
-	movl %eax,64(%edi)
-	movl %ebx,68(%edi)
-	movl 60(%edi),%eax
-	movl 64(%edi),%ebx
-	xorl 40(%edi),%eax
-	xorl 44(%edi),%ebx
-	xorl 16(%edi),%eax
-	xorl 20(%edi),%ebx
-	xorl 8(%edi),%eax
-	xorl 12(%edi),%ebx
-	roll $1,%eax
-	roll $1,%ebx
-	movl %eax,72(%edi)
-	movl %ebx,76(%edi)
-	addl $16,%edi
-	decl %ecx
-	jnz LOCAL(2)
-
-	movl $PARAM_DATA,%edi
-
-	movl (%ebp),%eax
-LOCAL(01_20):
-	subround1  4(%ebp), 8(%ebp), 12(%ebp), 16(%ebp), 0
-	subround1   (%ebp),   %ebx ,  8(%ebp), 12(%ebp), 4
-	subround1 16(%ebp),   %ebx ,  4(%ebp),  8(%ebp), 8
-	subround1 12(%ebp),   %ebx ,   (%ebp),  4(%ebp), 12
-	subround1  8(%ebp),   %ebx , 16(%ebp),   (%ebp), 16
-	addl $20,%edi
-	subround1  4(%ebp),   %ebx , 12(%ebp), 16(%ebp), 0
-	subround1   (%ebp),   %ebx ,  8(%ebp), 12(%ebp), 4
-	subround1 16(%ebp),   %ebx ,  4(%ebp),  8(%ebp), 8
-	subround1 12(%ebp),   %ebx ,   (%ebp),  4(%ebp), 12
-	subround1  8(%ebp),   %ebx , 16(%ebp),   (%ebp), 16
-	addl $20,%edi
-	subround1  4(%ebp),   %ebx , 12(%ebp), 16(%ebp), 0
-	subround1   (%ebp),   %ebx ,  8(%ebp), 12(%ebp), 4
-	subround1 16(%ebp),   %ebx ,  4(%ebp),  8(%ebp), 8
-	subround1 12(%ebp),   %ebx ,   (%ebp),  4(%ebp), 12
-	subround1  8(%ebp),   %ebx , 16(%ebp),   (%ebp), 16
-	addl $20,%edi
-	subround1  4(%ebp),   %ebx , 12(%ebp), 16(%ebp), 0
-	subround1   (%ebp),   %ebx ,  8(%ebp), 12(%ebp), 4
-	subround1 16(%ebp),   %ebx ,  4(%ebp),  8(%ebp), 8
-	subround1 12(%ebp),   %ebx ,   (%ebp),  4(%ebp), 12
-	subround1  8(%ebp),   %ebx , 16(%ebp),   (%ebp), 16
-	addl $20,%edi
-
-LOCAL(21_40):
-	subround2  4(%ebp),   %ebx , 12(%ebp), 16(%ebp), 0
-	subround2   (%ebp),   %ebx ,  8(%ebp), 12(%ebp), 4
-	subround2 16(%ebp),   %ebx ,  4(%ebp),  8(%ebp), 8
-	subround2 12(%ebp),   %ebx ,   (%ebp),  4(%ebp), 12
-	subround2  8(%ebp),   %ebx , 16(%ebp),   (%ebp), 16
-	addl $20,%edi
-	subround2  4(%ebp),   %ebx , 12(%ebp), 16(%ebp), 0
-	subround2   (%ebp),   %ebx ,  8(%ebp), 12(%ebp), 4
-	subround2 16(%ebp),   %ebx ,  4(%ebp),  8(%ebp), 8
-	subround2 12(%ebp),   %ebx ,   (%ebp),  4(%ebp), 12
-	subround2  8(%ebp),   %ebx , 16(%ebp),   (%ebp), 16
-	addl $20,%edi
-	subround2  4(%ebp),   %ebx , 12(%ebp), 16(%ebp), 0
-	subround2   (%ebp),   %ebx ,  8(%ebp), 12(%ebp), 4
-	subround2 16(%ebp),   %ebx ,  4(%ebp),  8(%ebp), 8
-	subround2 12(%ebp),   %ebx ,   (%ebp),  4(%ebp), 12
-	subround2  8(%ebp),   %ebx , 16(%ebp),   (%ebp), 16
-	addl $20,%edi
-	subround2  4(%ebp),   %ebx , 12(%ebp), 16(%ebp), 0
-	subround2   (%ebp),   %ebx ,  8(%ebp), 12(%ebp), 4
-	subround2 16(%ebp),   %ebx ,  4(%ebp),  8(%ebp), 8
-	subround2 12(%ebp),   %ebx ,   (%ebp),  4(%ebp), 12
-	subround2  8(%ebp),   %ebx , 16(%ebp),   (%ebp), 16
-	addl $20,%edi
-
-LOCAL(41_60):
-	subround3  4(%ebp),   %ebx , 12(%ebp), 16(%ebp), 0
-	subround3   (%ebp),   %ebx ,  8(%ebp), 12(%ebp), 4
-	subround3 16(%ebp),   %ebx ,  4(%ebp),  8(%ebp), 8
-	subround3 12(%ebp),   %ebx ,   (%ebp),  4(%ebp), 12
-	subround3  8(%ebp),   %ebx , 16(%ebp),   (%ebp), 16
-	addl $20,%edi
-	subround3  4(%ebp),   %ebx , 12(%ebp), 16(%ebp), 0
-	subround3   (%ebp),   %ebx ,  8(%ebp), 12(%ebp), 4
-	subround3 16(%ebp),   %ebx ,  4(%ebp),  8(%ebp), 8
-	subround3 12(%ebp),   %ebx ,   (%ebp),  4(%ebp), 12
-	subround3  8(%ebp),   %ebx , 16(%ebp),   (%ebp), 16
-	addl $20,%edi
-	subround3  4(%ebp),   %ebx , 12(%ebp), 16(%ebp), 0
-	subround3   (%ebp),   %ebx ,  8(%ebp), 12(%ebp), 4
-	subround3 16(%ebp),   %ebx ,  4(%ebp),  8(%ebp), 8
-	subround3 12(%ebp),   %ebx ,   (%ebp),  4(%ebp), 12
-	subround3  8(%ebp),   %ebx , 16(%ebp),   (%ebp), 16
-	addl $20,%edi
-	subround3  4(%ebp),   %ebx , 12(%ebp), 16(%ebp), 0
-	subround3   (%ebp),   %ebx ,  8(%ebp), 12(%ebp), 4
-	subround3 16(%ebp),   %ebx ,  4(%ebp),  8(%ebp), 8
-	subround3 12(%ebp),   %ebx ,   (%ebp),  4(%ebp), 12
-	subround3  8(%ebp),   %ebx , 16(%ebp),   (%ebp), 16
-	addl $20,%edi
-
-LOCAL(61_80):
-	subround4  4(%ebp),   %ebx , 12(%ebp), 16(%ebp), 0
-	subround4   (%ebp),   %ebx ,  8(%ebp), 12(%ebp), 4
-	subround4 16(%ebp),   %ebx ,  4(%ebp),  8(%ebp), 8
-	subround4 12(%ebp),   %ebx ,   (%ebp),  4(%ebp), 12
-	subround4  8(%ebp),   %ebx , 16(%ebp),   (%ebp), 16
-	addl $20,%edi
-	subround4  4(%ebp),   %ebx , 12(%ebp), 16(%ebp), 0
-	subround4   (%ebp),   %ebx ,  8(%ebp), 12(%ebp), 4
-	subround4 16(%ebp),   %ebx ,  4(%ebp),  8(%ebp), 8
-	subround4 12(%ebp),   %ebx ,   (%ebp),  4(%ebp), 12
-	subround4  8(%ebp),   %ebx , 16(%ebp),   (%ebp), 16
-	addl $20,%edi
-	subround4  4(%ebp),   %ebx , 12(%ebp), 16(%ebp), 0
-	subround4   (%ebp),   %ebx ,  8(%ebp), 12(%ebp), 4
-	subround4 16(%ebp),   %ebx ,  4(%ebp),  8(%ebp), 8
-	subround4 12(%ebp),   %ebx ,   (%ebp),  4(%ebp), 12
-	subround4  8(%ebp),   %ebx , 16(%ebp),   (%ebp), 16
-	addl $20,%edi
-	subround4  4(%ebp),   %ebx , 12(%ebp), 16(%ebp), 0
-	subround4   (%ebp),   %ebx ,  8(%ebp), 12(%ebp), 4
-	subround4 16(%ebp),   %ebx ,  4(%ebp),  8(%ebp), 8
-	subround4 12(%ebp),   %ebx ,   (%ebp),  4(%ebp), 12
-	subround4  8(%ebp),   %ebx , 16(%ebp),   (%ebp), 16
-	/* addl $20,%edi */
-
-	movl $4,%ecx
-
-	.p2align 2
-LOCAL(3):
-	movl (%ebp,%ecx,4),%eax
-	addl %eax,(%esi,%ecx,4)
-	decl %ecx
-	jns LOCAL(3)
-
-	addl $20,%esp
-	popl %ebp
-	popl %ebx
-	popl %esi
-	popl %edi
-	ret
-C_FUNCTION_END(sha1Process, LOCAL(sha1Process_size))
diff --git a/beecrypt/gas/sha1opt.i586.m4 b/beecrypt/gas/sha1opt.i586.m4
new file mode 100644
index 000000000..a8d8ec6b5
--- /dev/null
+++ b/beecrypt/gas/sha1opt.i586.m4
@@ -0,0 +1,280 @@
+dnl  sha1opt.i586.m4
+dnl
+dnl  Copyright (c) 2003 Bob Deblier
+dnl 
+dnl  Author: Bob Deblier <bob.deblier@pandora.be>
+dnl 
+dnl  This library is free software; you can redistribute it and/or
+dnl  modify it under the terms of the GNU Lesser General Public
+dnl  License as published by the Free Software Foundation; either
+dnl  version 2.1 of the License, or (at your option) any later version.
+dnl 
+dnl  This library 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 GNU
+dnl  Lesser General Public License for more details.
+dnl 
+dnl  You should have received a copy of the GNU Lesser General Public
+dnl  License along with this library; if not, write to the Free Software
+dnl  Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
+
+include(config.m4)
+include(ASM_SRCDIR/x86.m4)
+
+	.equ	K00,	0x5a827999
+	.equ	K20,	0x6ed9eba1
+	.equ	K40,	0x8f1bbcdc
+	.equ	K60,	0xca62c1d6
+
+	.equ	PARAM_H,		0
+	.equ	PARAM_DATA,		20
+
+define(`subround1',`
+	movl $2,%ecx
+	movl $1,%ebx
+	movl $3,%edx
+	roll `$'5,%eax
+	xorl %edx,%ecx
+	addl $4,%eax
+	andl %ebx,%ecx
+	addl `$'K00,%eax
+	rorl `$'2,%ebx
+	addl $5(%esi,%edi),%eax
+	xorl %edx,%ecx
+	movl %ebx,$1
+	addl %ecx,%eax
+	movl %eax,$4
+')
+
+define(`subround2',`
+	movl $2,%ecx
+	movl $1,%ebx
+	roll `$'5,%eax
+	xorl %ebx,%ecx
+	addl $4,%eax
+	xorl $3,%ecx
+	addl `$'K20,%eax
+	rorl `$'2,%ebx
+	addl $5(%esi,%edi),%eax
+	movl %ebx,$1
+	addl %ecx,%eax
+	movl %eax,$4
+')
+
+define(`subround3',`
+	movl $2,%ecx
+	roll `$'5,%eax
+	movl $1,%ebx
+	movl %ecx,%edx
+	addl $4,%eax
+	orl %ebx,%ecx
+	andl %ebx,%edx
+	andl $3,%ecx
+	addl `$'K40,%eax
+	orl %edx,%ecx
+	addl $5(%esi,%edi),%eax
+	rorl `$'2,%ebx
+	addl %ecx,%eax
+	movl %ebx,$1
+	movl %eax,$4
+')
+
+define(`subround4',`
+	movl $2,%ecx
+	movl $1,%ebx
+	roll `$'5,%eax
+	xorl %ebx,%ecx
+	addl $4,%eax
+	xorl $3,%ecx
+	addl `$'K60,%eax
+	rorl `$'2,%ebx
+	addl $5(%esi,%edi),%eax
+	movl %ebx,$1
+	addl %ecx,%eax
+	movl %eax,$4
+')
+
+C_FUNCTION_BEGIN(sha1Process)
+	pushl %edi
+	pushl %esi
+	pushl %ebx
+	pushl %ebp
+
+	movl 20(%esp),%esi
+	subl `$'20,%esp
+	leal PARAM_DATA(%esi),%edi
+	movl %esp,%ebp
+
+	movl `$'4,%ecx
+LOCAL(0):
+	movl (%esi,%ecx,4),%edx
+	movl %edx,(%ebp,%ecx,4)
+	decl %ecx
+	jns LOCAL(0)
+
+	movl `$'15,%ecx
+	xorl %eax,%eax
+
+	.align 4
+LOCAL(1):
+	movl (%edi,%ecx,4),%edx
+	bswap %edx
+	mov %edx,(%edi,%ecx,4)
+	decl %ecx
+	jns LOCAL(1)
+
+	leal PARAM_DATA(%esi),%edi
+	movl `$'16,%ecx
+
+	.align 4
+LOCAL(2):
+	movl 52(%edi),%eax
+	movl 56(%edi),%ebx
+	xorl 32(%edi),%eax
+	xorl 36(%edi),%ebx
+	xorl 8(%edi),%eax
+	xorl 12(%edi),%ebx
+	xorl (%edi),%eax
+	xorl 4(%edi),%ebx
+	roll `$'1,%eax
+	roll `$'1,%ebx
+	movl %eax,64(%edi)
+	movl %ebx,68(%edi)
+	movl 60(%edi),%eax
+	movl 64(%edi),%ebx
+	xorl 40(%edi),%eax
+	xorl 44(%edi),%ebx
+	xorl 16(%edi),%eax
+	xorl 20(%edi),%ebx
+	xorl 8(%edi),%eax
+	xorl 12(%edi),%ebx
+	roll `$'1,%eax
+	roll `$'1,%ebx
+	movl %eax,72(%edi)
+	movl %ebx,76(%edi)
+	addl `$'16,%edi
+	decl %ecx
+	jnz LOCAL(2)
+
+	movl `$'PARAM_DATA,%edi
+
+	movl (%ebp),%eax
+LOCAL(01_20):
+	subround1( 4(%ebp), 8(%ebp), 12(%ebp), 16(%ebp), 0)
+	subround1(  (%ebp),   %ebx ,  8(%ebp), 12(%ebp), 4)
+	subround1(16(%ebp),   %ebx ,  4(%ebp),  8(%ebp), 8)
+	subround1(12(%ebp),   %ebx ,   (%ebp),  4(%ebp), 12)
+	subround1( 8(%ebp),   %ebx , 16(%ebp),   (%ebp), 16)
+	addl `$'20,%edi
+	subround1( 4(%ebp),   %ebx , 12(%ebp), 16(%ebp), 0)
+	subround1(  (%ebp),   %ebx ,  8(%ebp), 12(%ebp), 4)
+	subround1(16(%ebp),   %ebx ,  4(%ebp),  8(%ebp), 8)
+	subround1(12(%ebp),   %ebx ,   (%ebp),  4(%ebp), 12)
+	subround1( 8(%ebp),   %ebx , 16(%ebp),   (%ebp), 16)
+	addl `$'20,%edi
+	subround1( 4(%ebp),   %ebx , 12(%ebp), 16(%ebp), 0)
+	subround1(  (%ebp),   %ebx ,  8(%ebp), 12(%ebp), 4)
+	subround1(16(%ebp),   %ebx ,  4(%ebp),  8(%ebp), 8)
+	subround1(12(%ebp),   %ebx ,   (%ebp),  4(%ebp), 12)
+	subround1( 8(%ebp),   %ebx , 16(%ebp),   (%ebp), 16)
+	addl `$'20,%edi
+	subround1( 4(%ebp),   %ebx , 12(%ebp), 16(%ebp), 0)
+	subround1(  (%ebp),   %ebx ,  8(%ebp), 12(%ebp), 4)
+	subround1(16(%ebp),   %ebx ,  4(%ebp),  8(%ebp), 8)
+	subround1(12(%ebp),   %ebx ,   (%ebp),  4(%ebp), 12)
+	subround1( 8(%ebp),   %ebx , 16(%ebp),   (%ebp), 16)
+	addl `$'20,%edi
+
+LOCAL(21_40):
+	subround2( 4(%ebp),   %ebx , 12(%ebp), 16(%ebp), 0)
+	subround2(  (%ebp),   %ebx ,  8(%ebp), 12(%ebp), 4)
+	subround2(16(%ebp),   %ebx ,  4(%ebp),  8(%ebp), 8)
+	subround2(12(%ebp),   %ebx ,   (%ebp),  4(%ebp), 12)
+	subround2( 8(%ebp),   %ebx , 16(%ebp),   (%ebp), 16)
+	addl `$'20,%edi
+	subround2( 4(%ebp),   %ebx , 12(%ebp), 16(%ebp), 0)
+	subround2(  (%ebp),   %ebx ,  8(%ebp), 12(%ebp), 4)
+	subround2(16(%ebp),   %ebx ,  4(%ebp),  8(%ebp), 8)
+	subround2(12(%ebp),   %ebx ,   (%ebp),  4(%ebp), 12)
+	subround2( 8(%ebp),   %ebx , 16(%ebp),   (%ebp), 16)
+	addl `$'20,%edi
+	subround2( 4(%ebp),   %ebx , 12(%ebp), 16(%ebp), 0)
+	subround2(  (%ebp),   %ebx ,  8(%ebp), 12(%ebp), 4)
+	subround2(16(%ebp),   %ebx ,  4(%ebp),  8(%ebp), 8)
+	subround2(12(%ebp),   %ebx ,   (%ebp),  4(%ebp), 12)
+	subround2( 8(%ebp),   %ebx , 16(%ebp),   (%ebp), 16)
+	addl `$'20,%edi
+	subround2( 4(%ebp),   %ebx , 12(%ebp), 16(%ebp), 0)
+	subround2(  (%ebp),   %ebx ,  8(%ebp), 12(%ebp), 4)
+	subround2(16(%ebp),   %ebx ,  4(%ebp),  8(%ebp), 8)
+	subround2(12(%ebp),   %ebx ,   (%ebp),  4(%ebp), 12)
+	subround2( 8(%ebp),   %ebx , 16(%ebp),   (%ebp), 16)
+	addl `$'20,%edi
+
+LOCAL(41_60):
+	subround3( 4(%ebp),   %ebx , 12(%ebp), 16(%ebp), 0)
+	subround3(  (%ebp),   %ebx ,  8(%ebp), 12(%ebp), 4)
+	subround3(16(%ebp),   %ebx ,  4(%ebp),  8(%ebp), 8)
+	subround3(12(%ebp),   %ebx ,   (%ebp),  4(%ebp), 12)
+	subround3( 8(%ebp),   %ebx , 16(%ebp),   (%ebp), 16)
+	addl `$'20,%edi
+	subround3( 4(%ebp),   %ebx , 12(%ebp), 16(%ebp), 0)
+	subround3(  (%ebp),   %ebx ,  8(%ebp), 12(%ebp), 4)
+	subround3(16(%ebp),   %ebx ,  4(%ebp),  8(%ebp), 8)
+	subround3(12(%ebp),   %ebx ,   (%ebp),  4(%ebp), 12)
+	subround3( 8(%ebp),   %ebx , 16(%ebp),   (%ebp), 16)
+	addl `$'20,%edi
+	subround3( 4(%ebp),   %ebx , 12(%ebp), 16(%ebp), 0)
+	subround3(  (%ebp),   %ebx ,  8(%ebp), 12(%ebp), 4)
+	subround3(16(%ebp),   %ebx ,  4(%ebp),  8(%ebp), 8)
+	subround3(12(%ebp),   %ebx ,   (%ebp),  4(%ebp), 12)
+	subround3( 8(%ebp),   %ebx , 16(%ebp),   (%ebp), 16)
+	addl `$'20,%edi
+	subround3( 4(%ebp),   %ebx , 12(%ebp), 16(%ebp), 0)
+	subround3(  (%ebp),   %ebx ,  8(%ebp), 12(%ebp), 4)
+	subround3(16(%ebp),   %ebx ,  4(%ebp),  8(%ebp), 8)
+	subround3(12(%ebp),   %ebx ,   (%ebp),  4(%ebp), 12)
+	subround3( 8(%ebp),   %ebx , 16(%ebp),   (%ebp), 16)
+	addl `$'20,%edi
+
+LOCAL(61_80):
+	subround4( 4(%ebp),   %ebx , 12(%ebp), 16(%ebp), 0)
+	subround4(  (%ebp),   %ebx ,  8(%ebp), 12(%ebp), 4)
+	subround4(16(%ebp),   %ebx ,  4(%ebp),  8(%ebp), 8)
+	subround4(12(%ebp),   %ebx ,   (%ebp),  4(%ebp), 12)
+	subround4( 8(%ebp),   %ebx , 16(%ebp),   (%ebp), 16)
+	addl `$'20,%edi
+	subround4( 4(%ebp),   %ebx , 12(%ebp), 16(%ebp), 0)
+	subround4(  (%ebp),   %ebx ,  8(%ebp), 12(%ebp), 4)
+	subround4(16(%ebp),   %ebx ,  4(%ebp),  8(%ebp), 8)
+	subround4(12(%ebp),   %ebx ,   (%ebp),  4(%ebp), 12)
+	subround4( 8(%ebp),   %ebx , 16(%ebp),   (%ebp), 16)
+	addl `$'20,%edi
+	subround4( 4(%ebp),   %ebx , 12(%ebp), 16(%ebp), 0)
+	subround4(  (%ebp),   %ebx ,  8(%ebp), 12(%ebp), 4)
+	subround4(16(%ebp),   %ebx ,  4(%ebp),  8(%ebp), 8)
+	subround4(12(%ebp),   %ebx ,   (%ebp),  4(%ebp), 12)
+	subround4( 8(%ebp),   %ebx , 16(%ebp),   (%ebp), 16)
+	addl `$'20,%edi
+	subround4( 4(%ebp),   %ebx , 12(%ebp), 16(%ebp), 0)
+	subround4(  (%ebp),   %ebx ,  8(%ebp), 12(%ebp), 4)
+	subround4(16(%ebp),   %ebx ,  4(%ebp),  8(%ebp), 8)
+	subround4(12(%ebp),   %ebx ,   (%ebp),  4(%ebp), 12)
+	subround4( 8(%ebp),   %ebx , 16(%ebp),   (%ebp), 16)
+
+	movl `$'4,%ecx
+
+	.align 4
+LOCAL(3):
+	movl (%ebp,%ecx,4),%eax
+	addl %eax,(%esi,%ecx,4)
+	decl %ecx
+	jns LOCAL(3)
+
+	addl `$'20,%esp
+	popl %ebp
+	popl %ebx
+	popl %esi
+	popl %edi
+	ret
+C_FUNCTION_END(sha1Process)
diff --git a/beecrypt/gas/sha1opt.powerpc.S b/beecrypt/gas/sha1opt.powerpc.S
deleted file mode 100644
index f9dc73a9e..000000000
--- a/beecrypt/gas/sha1opt.powerpc.S
+++ /dev/null
@@ -1,287 +0,0 @@
-/*
- * fips180opt.powerpc.S
- *
- * Assembler optimized SHA-1 routines for PowerPC processors
- *
- * Warning: this code is incomplete and only contains a rough prototype!
- *
- * Compile target is GNU Assembler
- *
- * Copyright (c) 2000, 2001 Virtual Unlimited B.V.
- *
- * Author: Bob Deblier <bob@virtualunlimited.com>
- *
- * This library is free software; you can redistribute it and/or
- * modify it under the terms of the GNU Lesser General Public
- * License as published by the Free Software Foundation; either
- * version 2.1 of the License, or (at your option) any later version.
- *
- * This library 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
- * Lesser General Public License for more details.
- *
- * You should have received a copy of the GNU Lesser General Public
- * License along with this library; if not, write to the Free Software
- * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
- *
- */
-
-#include "beecrypt.gas.h"
-
-	.file "fips180opt.powerpc.S"
-
-	.text
-
-#if DARWIN
-# define reg0	r0
-# define reg3	r3
-# define reg4	r4
-# define reg5	r5
-# define reg6	r6
-# define reg7	r7
-# define reg8	r8
-# define reg9	r9
-# define reg26	r26
-# define reg27	r27
-# define reg28	r28
-# define reg29	r29
-# define reg30	r30
-# define reg31	r31
-#else
-# define reg0	%r0
-# define reg3	%r3
-# define reg4	%r4
-# define reg5	%r5
-# define reg6	%r6
-# define reg7	%r7
-# define reg8	%r8
-# define reg9	%r9
-# define reg26	%r26
-# define reg27	%r27
-# define reg28	%r28
-# define reg29	%r29
-# define reg30	%r30
-# define reg31	%r31
-#endif
-
-#define K00	0x5a827999
-#define K20	0x6ed9eba1
-#define K40	0x8f1bbcdc
-#define K60	0xca62c1d6
-
-#define PARAM_H		0
-#define PARAM_DATA	20
-
-/* sha1Param: param in reg3 */
-
-	.macro	subround1 a b c d e w
-	lwzu reg7,4(\w)
-	rotlwi	reg5,\a,5
-	dbct r0,\w
-	xor reg6,\c,\d
-	add \e,\e,K00
-	and reg6,reg6,\b
-	add \e,\e,reg7
-	xor reg6,reg6,\d
-	add \e,\e,reg5
-	rotrwi \b,\b,2
-	add \e,\e,reg6
-	.endm
-
-	.macro	subround2 a b c d e w
-	lwzu reg7,4(\w)
-	rotlwi reg5,\a,5
-	dbct r0,\w
-	add \e,\e,K20
-	xor reg6,\b,\c
-	add \e,\e,reg5
-	xor reg6,reg6,\d
-	add \e,\e,reg7
-	rotrwi \b,\b,2
-	add \e,\e,reg6
-	.endm
-
-	.macro	subround3 a b c d e w
-	lwzu reg7,4(\w)
-	rotlwi reg5,\a,5
-	dbct r0,\w
-	xor reg6,\b,\c
-	add \e,\e,reg5
-	and reg6,reg6,\d
-	add \e,\e,K40
-	and reg5,\b,\c
-	add \e,\e,reg7
-	or reg6,reg6,reg5
-	rotrwi \b,\b,2
-	add \e,\e,reg6
-	.endm
-
-	.macro	subround4 a b c d e w
-	lwzu reg7,4(\w)
-	rotlwi reg5,\a,5
-	dbct r0,\w
-	add \e,\e,K60
-	xor reg6,\b,\c
-	add \e,\e,reg5
-	xor reg6,reg6,\d
-	add \e,\e,reg7
-	rotrwi \b,\b,2
-	add \e,\e,reg6
-	.endm
-
-C_FUNCTION_BEGIN(sha1Process)
-/* zero reg0 for general use */
-	li reg0,0
-/* for a,b,c,d,e use r26,r27,r28,r29,r30, for w use r31 */
-	
-/* we need to save registers before loading them */
-	stmw reg26,-24(reg1)
-/* load the frame pointer with parameter data, and hint cache */
-	addi reg31,reg3,PARAM_DATA
-	dbct reg31
-
-#if !WORDS_BIGENDIAN /* have to provide for PowerPC little-endian mode
-	/* loop of 16 entries */
-	li reg5,60
-	mtctr reg6
-.L00:
-	lwbrx reg6,reg31,reg5
-	stwx reg6,reg31,reg5
-	subi. reg5,reg5,4
-	bcge cr0,.L00
-	addi reg31,reg3,PARAM_DATA
-#endif
-
-/* do the initial mixing */
-	li reg8,64
-	addi reg26,reg3,PARAM_DATA+64-4
-	addi reg27,reg3,PARAM_DATA+64-3*4-4
-	addi reg28,reg3,PARAM_DATA+64-8*4-4
-	addi reg29,reg3,PARAM_DATA+64-14*4-4
-	addi reg30,reg3,PARAM_DATA+64-16*4-4
-	mtctr reg8
-
-.L10:
-	lwzu reg5,4(reg27)
-	lwzu reg6,4(reg28)
-	lwzu reg7,4(reg29)
-	lwzu reg8,4(reg30)
-	xor reg5,reg5,reg6
-	xor reg7,reg7,reg8
-	xor reg5,reg5,reg7
-	stwu reg5,4(reg26)
-	bdnz .L10
-
-	lwz reg26,PARAM_H   (reg3)
-	lwz reg27,PARAM_H+4 (reg3)
-	lwz reg28,PARAM_H+8 (reg3)
-	lwz reg29,PARAM_H+12(reg3)
-	lwz reg30,PARAM_H+16(reg3)
-
-	subround1 reg26,reg27,reg28,reg29,reg30,reg31
-	subround1 reg30,reg26,reg27,reg28,reg29,reg31
-	subround1 reg29,reg30,reg26,reg27,reg28,reg31
-	subround1 reg28,reg29,reg30,reg26,reg27,reg31
-	subround1 reg27,reg28,reg29,reg30,reg26,reg31
-	subround1 reg26,reg27,reg28,reg29,reg30,reg31
-	subround1 reg30,reg26,reg27,reg28,reg29,reg31
-	subround1 reg29,reg30,reg26,reg27,reg28,reg31
-	subround1 reg28,reg29,reg30,reg26,reg27,reg31
-	subround1 reg27,reg28,reg29,reg30,reg26,reg31
-	subround1 reg26,reg27,reg28,reg29,reg30,reg31
-	subround1 reg30,reg26,reg27,reg28,reg29,reg31
-	subround1 reg29,reg30,reg26,reg27,reg28,reg31
-	subround1 reg28,reg29,reg30,reg26,reg27,reg31
-	subround1 reg27,reg28,reg29,reg30,reg26,reg31
-	subround1 reg26,reg27,reg28,reg29,reg30,reg31
-	subround1 reg30,reg26,reg27,reg28,reg29,reg31
-	subround1 reg29,reg30,reg26,reg27,reg28,reg31
-	subround1 reg28,reg29,reg30,reg26,reg27,reg31
-	subround1 reg27,reg28,reg29,reg30,reg26,reg31
-
-	subround2 reg26,reg27,reg28,reg29,reg30,reg31
-	subround2 reg30,reg26,reg27,reg28,reg29,reg31
-	subround2 reg29,reg30,reg26,reg27,reg28,reg31
-	subround2 reg28,reg29,reg30,reg26,reg27,reg31
-	subround2 reg27,reg28,reg29,reg30,reg26,reg31
-	subround2 reg26,reg27,reg28,reg29,reg30,reg31
-	subround2 reg30,reg26,reg27,reg28,reg29,reg31
-	subround2 reg29,reg30,reg26,reg27,reg28,reg31
-	subround2 reg28,reg29,reg30,reg26,reg27,reg31
-	subround2 reg27,reg28,reg29,reg30,reg26,reg31
-	subround2 reg26,reg27,reg28,reg29,reg30,reg31
-	subround2 reg30,reg26,reg27,reg28,reg29,reg31
-	subround2 reg29,reg30,reg26,reg27,reg28,reg31
-	subround2 reg28,reg29,reg30,reg26,reg27,reg31
-	subround2 reg27,reg28,reg29,reg30,reg26,reg31
-	subround2 reg26,reg27,reg28,reg29,reg30,reg31
-	subround2 reg30,reg26,reg27,reg28,reg29,reg31
-	subround2 reg29,reg30,reg26,reg27,reg28,reg31
-	subround2 reg28,reg29,reg30,reg26,reg27,reg31
-	subround2 reg27,reg28,reg29,reg30,reg26,reg31
-
-	subround3 reg26,reg27,reg28,reg29,reg30,reg31
-	subround3 reg30,reg26,reg27,reg28,reg29,reg31
-	subround3 reg29,reg30,reg26,reg27,reg28,reg31
-	subround3 reg28,reg29,reg30,reg26,reg27,reg31
-	subround3 reg27,reg28,reg29,reg30,reg26,reg31
-	subround3 reg26,reg27,reg28,reg29,reg30,reg31
-	subround3 reg30,reg26,reg27,reg28,reg29,reg31
-	subround3 reg29,reg30,reg26,reg27,reg28,reg31
-	subround3 reg28,reg29,reg30,reg26,reg27,reg31
-	subround3 reg27,reg28,reg29,reg30,reg26,reg31
-	subround3 reg26,reg27,reg28,reg29,reg30,reg31
-	subround3 reg30,reg26,reg27,reg28,reg29,reg31
-	subround3 reg29,reg30,reg26,reg27,reg28,reg31
-	subround3 reg28,reg29,reg30,reg26,reg27,reg31
-	subround3 reg27,reg28,reg29,reg30,reg26,reg31
-	subround3 reg26,reg27,reg28,reg29,reg30,reg31
-	subround3 reg30,reg26,reg27,reg28,reg29,reg31
-	subround3 reg29,reg30,reg26,reg27,reg28,reg31
-	subround3 reg28,reg29,reg30,reg26,reg27,reg31
-	subround3 reg27,reg28,reg29,reg30,reg26,reg31
-
-	subround4 reg26,reg27,reg28,reg29,reg30,reg31
-	subround4 reg30,reg26,reg27,reg28,reg29,reg31
-	subround4 reg29,reg30,reg26,reg27,reg28,reg31
-	subround4 reg28,reg29,reg30,reg26,reg27,reg31
-	subround4 reg27,reg28,reg29,reg30,reg26,reg31
-	subround4 reg26,reg27,reg28,reg29,reg30,reg31
-	subround4 reg30,reg26,reg27,reg28,reg29,reg31
-	subround4 reg29,reg30,reg26,reg27,reg28,reg31
-	subround4 reg28,reg29,reg30,reg26,reg27,reg31
-	subround4 reg27,reg28,reg29,reg30,reg26,reg31
-	subround4 reg26,reg27,reg28,reg29,reg30,reg31
-	subround4 reg30,reg26,reg27,reg28,reg29,reg31
-	subround4 reg29,reg30,reg26,reg27,reg28,reg31
-	subround4 reg28,reg29,reg30,reg26,reg27,reg31
-	subround4 reg27,reg28,reg29,reg30,reg26,reg31
-	subround4 reg26,reg27,reg28,reg29,reg30,reg31
-	subround4 reg30,reg26,reg27,reg28,reg29,reg31
-	subround4 reg29,reg30,reg26,reg27,reg28,reg31
-	subround4 reg28,reg29,reg30,reg26,reg27,reg31
-	subround4 reg27,reg28,reg29,reg30,reg26,reg31
-
-/* then store the five values into registers */
-	lwz reg5,PARAM_H   (reg3)
-	lwz reg6,PARAM_H+4 (reg3)
-	lwz reg7,PARAM_H+8 (reg3)
-	lwz reg8,PARAM_H+12(reg3)
-	lwz reg9,PARAM_H+16(reg3)
-	add reg26,reg5,reg26
-	add reg27,reg5,reg27
-	add reg28,reg5,reg28
-	add reg29,reg5,reg29
-	add reg30,reg5,reg30
-	stw reg26,PARAM_H   (reg3)
-	stw reg27,PARAM_H+4 (reg3)
-	stw reg28,PARAM_H+8 (reg3)
-	stw reg29,PARAM_H+12(reg3)
-	stw reg30,PARAM_H+16(reg3)
-
-/* finally, restore registers */
-	lmw reg26,-24(reg1)
-/* and return */
-	blr
-C_FUNCION_END(sha1Process, .Lsha1Process_size)
diff --git a/beecrypt/gas/sparc.m4 b/beecrypt/gas/sparc.m4
new file mode 100644
index 000000000..e735600c3
--- /dev/null
+++ b/beecrypt/gas/sparc.m4
@@ -0,0 +1,30 @@
+dnl  sparc.m4
+dnl
+dnl  Copyright (c) 2003 Bob Deblier
+dnl 
+dnl  Author: Bob Deblier <bob.deblier@pandora.be>
+dnl 
+dnl  This library is free software; you can redistribute it and/or
+dnl  modify it under the terms of the GNU Lesser General Public
+dnl  License as published by the Free Software Foundation; either
+dnl  version 2.1 of the License, or (at your option) any later version.
+dnl 
+dnl  This library 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 GNU
+dnl  Lesser General Public License for more details.
+dnl 
+dnl  You should have received a copy of the GNU Lesser General Public
+dnl  License along with this library; if not, write to the Free Software
+dnl  Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
+
+ifelse(substr(ASM_OS,0,7),solaris,`
+undefine(`C_FUNCTION_BEGIN')
+define(C_FUNCTION_BEGIN,`
+    TEXTSEG
+    GLOBL SYMNAME($1)
+SYMNAME($1):
+	.register %g2,#scratch
+	.register %g3,#scratch
+')
+')
diff --git a/beecrypt/gas/x86.m4 b/beecrypt/gas/x86.m4
new file mode 100644
index 000000000..131c94df2
--- /dev/null
+++ b/beecrypt/gas/x86.m4
@@ -0,0 +1,23 @@
+dnl  x86.m4
+dnl
+dnl  Copyright (c) 2003 Bob Deblier
+dnl 
+dnl  Author: Bob Deblier <bob.deblier@pandora.be>
+dnl 
+dnl  This library is free software; you can redistribute it and/or
+dnl  modify it under the terms of the GNU Lesser General Public
+dnl  License as published by the Free Software Foundation; either
+dnl  version 2.1 of the License, or (at your option) any later version.
+dnl 
+dnl  This library 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 GNU
+dnl  Lesser General Public License for more details.
+dnl 
+dnl  You should have received a copy of the GNU Lesser General Public
+dnl  License along with this library; if not, write to the Free Software
+dnl  Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
+
+ifelse(ASM_ARCH,pentium4,`
+  define(`USE_SSE2')
+')