[thirdparty/glibc.git] / sysdeps / sparc / sparc32 / sparcv8 / addmul_1.S

! SPARC v8 __mpn_addmul_1 -- Multiply a limb vector with a limb and
! add the result to a second limb vector.

! Copyright (C) 1992-2015 Free Software Foundation, Inc.

! This file is part of the GNU MP Library.

! The GNU MP 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.

! The GNU MP 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 the GNU MP Library; see the file COPYING.LIB.  If not,
! see <http://www.gnu.org/licenses/>.


! INPUT PARAMETERS
! res_ptr	o0
! s1_ptr	o1
! size		o2
! s2_limb	o3

#include <sysdep.h>

ENTRY(__mpn_addmul_1)
	ld	[%o1+0],%o4	! 1
	sll	%o2,4,%g1
	orcc	%g0,%g0,%g2
	mov	%o7,%g4			! Save return address register
	and	%g1,(4-1)<<4,%g1
1:	call	2f
	 add	%o7,3f-1b,%g3
2:	jmp	%g3+%g1
	 mov	%g4,%o7			! Restore return address register

	.align	4
3:
LOC(00):
	add	%o0,-4,%o0
	b	LOC(loop00)		/* 4, 8, 12, ... */
	 add	%o1,-4,%o1
	nop
LOC(01):
	b	LOC(loop01)		/* 1, 5, 9, ... */
	 nop
	nop
	nop
LOC(10):
	add	%o0,-12,%o0	/* 2, 6, 10, ... */
	b	LOC(loop10)
	 add	%o1,4,%o1
	nop
LOC(11):
	add	%o0,-8,%o0	/* 3, 7, 11, ... */
	b	LOC(loop11)
	 add	%o1,-8,%o1
	nop

LOC(loop):
	addcc	%g3,%g2,%g3	! 1
	ld	[%o1+4],%o4	! 2
	rd	%y,%g2		! 1
	addx	%g0,%g2,%g2
	ld	[%o0+0],%g1	! 2
	addcc	%g1,%g3,%g3
	st	%g3,[%o0+0]	! 1
LOC(loop00):
	umul	%o4,%o3,%g3	! 2
	ld	[%o0+4],%g1	! 2
	addxcc	%g3,%g2,%g3	! 2
	ld	[%o1+8],%o4	! 3
	rd	%y,%g2		! 2
	addx	%g0,%g2,%g2
	nop
	addcc	%g1,%g3,%g3
	st	%g3,[%o0+4]	! 2
LOC(loop11):
	umul	%o4,%o3,%g3	! 3
	addxcc	%g3,%g2,%g3	! 3
	ld	[%o1+12],%o4	! 4
	rd	%y,%g2		! 3
	add	%o1,16,%o1
	addx	%g0,%g2,%g2
	ld	[%o0+8],%g1	! 2
	addcc	%g1,%g3,%g3
	st	%g3,[%o0+8]	! 3
LOC(loop10):
	umul	%o4,%o3,%g3	! 4
	addxcc	%g3,%g2,%g3	! 4
	ld	[%o1+0],%o4	! 1
	rd	%y,%g2		! 4
	addx	%g0,%g2,%g2
	ld	[%o0+12],%g1	! 2
	addcc	%g1,%g3,%g3
	st	%g3,[%o0+12]	! 4
	add	%o0,16,%o0
	addx	%g0,%g2,%g2
LOC(loop01):
	addcc	%o2,-4,%o2
	bg	LOC(loop)
	 umul	%o4,%o3,%g3	! 1

	addcc	%g3,%g2,%g3	! 4
	rd	%y,%g2		! 4
	addx	%g0,%g2,%g2
	ld	[%o0+0],%g1	! 2
	addcc	%g1,%g3,%g3
	st	%g3,[%o0+0]	! 4
	retl
	 addx	%g0,%g2,%o0

END(__mpn_addmul_1)
Commit	Line	Data
28f540f4 RM	1	! SPARC v8 __mpn_addmul_1 -- Multiply a limb vector with a limb and
	2	! add the result to a second limb vector.
	3
b168057a	4	! Copyright (C) 1992-2015 Free Software Foundation, Inc.
28f540f4 RM	5
	6	! This file is part of the GNU MP Library.
	7
	8	! The GNU MP Library is free software; you can redistribute it and/or modify
6d84f89a AJ	9	! it under the terms of the GNU Lesser General Public License as published by
6d84f89a AJ	10	! the Free Software Foundation; either version 2.1 of the License, or (at your
28f540f4 RM	11	! option) any later version.
	12
	13	! The GNU MP Library is distributed in the hope that it will be useful, but
	14	! WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY
6d84f89a	15	! or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public
28f540f4 RM	16	! License for more details.
28f540f4 RM	17
6d84f89a	18	! You should have received a copy of the GNU Lesser General Public License
59ba27a6 PE	19	! along with the GNU MP Library; see the file COPYING.LIB. If not,
59ba27a6 PE	20	! see <http://www.gnu.org/licenses/>.
28f540f4 RM	21
	22
	23	! INPUT PARAMETERS
	24	! res_ptr o0
	25	! s1_ptr o1
	26	! size o2
	27	! s2_limb o3
	28
f41c8091	29	#include <sysdep.h>
28f540f4	30
f41c8091	31	ENTRY(__mpn_addmul_1)
28f540f4	32	ld [%o1+0],%o4 ! 1
28f540f4	33	sll %o2,4,%g1
f41c8091	34	orcc %g0,%g0,%g2
8f5ca04b	35	mov %o7,%g4 ! Save return address register
f41c8091 UD	36	and %g1,(4-1)<<4,%g1
	37	1: call 2f
	38	add %o7,3f-1b,%g3
	39	2: jmp %g3+%g1
	40	mov %g4,%o7 ! Restore return address register
	41
	42	.align 4
	43	3:
	44	LOC(00):
	45	add %o0,-4,%o0
	46	b LOC(loop00) /* 4, 8, 12, ... */
	47	add %o1,-4,%o1
28f540f4	48	nop
f41c8091 UD	49	LOC(01):
	50	b LOC(loop01) /* 1, 5, 9, ... */
	51	nop
28f540f4 RM	52	nop
28f540f4 RM	53	nop
f41c8091 UD	54	LOC(10):
	55	add %o0,-12,%o0 /* 2, 6, 10, ... */
	56	b LOC(loop10)
	57	add %o1,4,%o1
28f540f4	58	nop
f41c8091 UD	59	LOC(11):
	60	add %o0,-8,%o0 /* 3, 7, 11, ... */
	61	b LOC(loop11)
	62	add %o1,-8,%o1
28f540f4 RM	63	nop
28f540f4 RM	64
f41c8091 UD	65	LOC(loop):
f41c8091 UD	66	addcc %g3,%g2,%g3 ! 1
28f540f4 RM	67	ld [%o1+4],%o4 ! 2
	68	rd %y,%g2 ! 1
	69	addx %g0,%g2,%g2
	70	ld [%o0+0],%g1 ! 2
	71	addcc %g1,%g3,%g3
	72	st %g3,[%o0+0] ! 1
f41c8091 UD	73	LOC(loop00):
f41c8091 UD	74	umul %o4,%o3,%g3 ! 2
28f540f4 RM	75	ld [%o0+4],%g1 ! 2
	76	addxcc %g3,%g2,%g3 ! 2
	77	ld [%o1+8],%o4 ! 3
	78	rd %y,%g2 ! 2
	79	addx %g0,%g2,%g2
	80	nop
	81	addcc %g1,%g3,%g3
	82	st %g3,[%o0+4] ! 2
f41c8091 UD	83	LOC(loop11):
f41c8091 UD	84	umul %o4,%o3,%g3 ! 3
28f540f4 RM	85	addxcc %g3,%g2,%g3 ! 3
	86	ld [%o1+12],%o4 ! 4
	87	rd %y,%g2 ! 3
	88	add %o1,16,%o1
	89	addx %g0,%g2,%g2
	90	ld [%o0+8],%g1 ! 2
	91	addcc %g1,%g3,%g3
	92	st %g3,[%o0+8] ! 3
f41c8091 UD	93	LOC(loop10):
f41c8091 UD	94	umul %o4,%o3,%g3 ! 4
28f540f4 RM	95	addxcc %g3,%g2,%g3 ! 4
	96	ld [%o1+0],%o4 ! 1
	97	rd %y,%g2 ! 4
	98	addx %g0,%g2,%g2
	99	ld [%o0+12],%g1 ! 2
	100	addcc %g1,%g3,%g3
	101	st %g3,[%o0+12] ! 4
	102	add %o0,16,%o0
	103	addx %g0,%g2,%g2
f41c8091 UD	104	LOC(loop01):
	105	addcc %o2,-4,%o2
	106	bg LOC(loop)
	107	umul %o4,%o3,%g3 ! 1
28f540f4 RM	108
	109	addcc %g3,%g2,%g3 ! 4
	110	rd %y,%g2 ! 4
	111	addx %g0,%g2,%g2
	112	ld [%o0+0],%g1 ! 2
	113	addcc %g1,%g3,%g3
	114	st %g3,[%o0+0] ! 4
28f540f4	115	retl
f41c8091	116	addx %g0,%g2,%o0
28f540f4	117
f41c8091	118	END(__mpn_addmul_1)