/* IBM POWER addmul_1 -- Multiply a limb vector with a limb and add
 *			 the result to a second limb vector.
 *
 *      Copyright (C) 1992, 1994, 1999, 2002 Free Software Foundation, Inc.
 *
 * This file is part of Libgcrypt.
 *
 * Libgcrypt 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.
 *
 * Libgcrypt 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 program; if not, write to the Free Software
 * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA
 */

#include "sysdep.h"
#include "asm-syntax.h"


/*
# INPUT PARAMETERS
# res_ptr	r3
# s1_ptr	r4
# size		r5
# s2_limb	r6

# The RS/6000 has no unsigned 32x32->64 bit multiplication instruction.  To
# obtain that operation, we have to use the 32x32->64 signed multiplication
# instruction, and add the appropriate compensation to the high limb of the
# result.  We add the multiplicand if the multiplier has its most significant
# bit set, and we add the multiplier if the multiplicand has its most
# significant bit set.	We need to preserve the carry flag between each
# iteration, so we have to compute the compensation carefully (the natural,
# srai+and doesn't work).  Since the POWER architecture has a branch unit
# we can branch in zero cycles, so that's how we perform the additions.
 */

	.toc
	.csect ._gcry_mpih_addmul_1[PR]
	.align 2
	.globl _gcry_mpih_addmul_1
	.globl ._gcry_mpih_addmul_1
	.csect _gcry_mpih_addmul_1[DS]
_gcry_mpih_addmul_1:
	.long ._gcry_mpih_addmul_1[PR], TOC[tc0], 0
	.csect ._gcry_mpih_addmul_1[PR]
._gcry_mpih_addmul_1:

	cal	3,-4(3)
	l	0,0(4)
	cmpi	0,6,0
	mtctr	5
	mul	9,0,6
	srai	7,0,31
	and	7,7,6
	mfmq	8
	cax	9,9,7
	l	7,4(3)
	a	8,8,7		# add res_limb
	blt	Lneg
Lpos:	bdz	Lend

Lploop: lu	0,4(4)
	stu	8,4(3)
	cmpi	0,0,0
	mul	10,0,6
	mfmq	0
	ae	8,0,9		# low limb + old_cy_limb + old cy
	l	7,4(3)
	aze	10,10		# propagate cy to new cy_limb
	a	8,8,7		# add res_limb
	bge	Lp0
	cax	10,10,6 	# adjust high limb for negative limb from s1
Lp0:	bdz	Lend0
	lu	0,4(4)
	stu	8,4(3)
	cmpi	0,0,0
	mul	9,0,6
	mfmq	0
	ae	8,0,10
	l	7,4(3)
	aze	9,9
	a	8,8,7
	bge	Lp1
	cax	9,9,6		# adjust high limb for negative limb from s1
Lp1:	bdn	Lploop

	b	Lend

Lneg:	cax	9,9,0
	bdz	Lend
Lnloop: lu	0,4(4)
	stu	8,4(3)
	cmpi	0,0,0
	mul	10,0,6
	mfmq	7
	ae	8,7,9
	l	7,4(3)
	ae	10,10,0 	# propagate cy to new cy_limb
	a	8,8,7		# add res_limb
	bge	Ln0
	cax	10,10,6 	# adjust high limb for negative limb from s1
Ln0:	bdz	Lend0
	lu	0,4(4)
	stu	8,4(3)
	cmpi	0,0,0
	mul	9,0,6
	mfmq	7
	ae	8,7,10
	l	7,4(3)
	ae	9,9,0		# propagate cy to new cy_limb
	a	8,8,7		# add res_limb
	bge	Ln1
	cax	9,9,6		# adjust high limb for negative limb from s1
Ln1:	bdn	Lnloop
	b	Lend

Lend0:	cal	9,0(10)
Lend:	st	8,4(3)
	aze	3,9
	br