1dnl  PowerPC-64 mpn_addmul_1 -- Multiply a limb vector with a limb and add
2dnl  the result to a second limb vector.
3
4dnl  Copyright 1999-2001, 2003, 2005 Free Software Foundation, Inc.
5
6dnl  This file is part of the GNU MP Library.
7dnl
8dnl  The GNU MP Library is free software; you can redistribute it and/or modify
9dnl  it under the terms of either:
10dnl
11dnl    * the GNU Lesser General Public License as published by the Free
12dnl      Software Foundation; either version 3 of the License, or (at your
13dnl      option) any later version.
14dnl
15dnl  or
16dnl
17dnl    * the GNU General Public License as published by the Free Software
18dnl      Foundation; either version 2 of the License, or (at your option) any
19dnl      later version.
20dnl
21dnl  or both in parallel, as here.
22dnl
23dnl  The GNU MP Library is distributed in the hope that it will be useful, but
24dnl  WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY
25dnl  or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License
26dnl  for more details.
27dnl
28dnl  You should have received copies of the GNU General Public License and the
29dnl  GNU Lesser General Public License along with the GNU MP Library.  If not,
30dnl  see https://www.gnu.org/licenses/.
31
32include(`../config.m4')
33
34C                   cycles/limb
35C POWER3/PPC630:     ?
36C POWER4/PPC970:     12.5
37
38C INPUT PARAMETERS
39C rp      r3
40C up      r4
41C n       r5
42C v       r6,r7  or  r7,r8
43
44ASM_START()
45PROLOGUE(mpn_addmul_1)
46
47ifdef(`BROKEN_LONGLONG_PARAM',
48`         rldimi    r8, r7, 32,0        C assemble vlimb from separate 32-bit arguments
49          mr        r6, r8
50',`
51          rldimi    r7, r6, 32,0        C assemble vlimb from separate 32-bit arguments
52          mr        r6, r7
53')
54          li        r7, 0               C cy_limb = 0
55          mtctr     r5
56          addic     r0, r0, 0
57          addi      r3, r3, -8
58          addi      r4, r4, -8
59
60L(oop):   ldu       r0, 8(r4)
61          mulld     r9, r0, r6
62          adde      r12, r9, r7         C add old high limb and new low limb
63          srdi      r5, r9, 32
64          srdi      r11, r7, 32
65          adde      r5, r5, r11         C add high limb parts, set cy
66          mulhdu    r7, r0, r6
67          addze     r7, r7
68          ld        r10, 8(r3)
69          addc      r9, r12, r10
70          srdi      r5, r12, 32
71          srdi      r11, r10, 32
72          adde      r5, r5, r11         C add high limb parts, set cy
73          stdu      r9, 8(r3)
74          bdnz      L(oop)
75
76          addze     r4, r7
77          srdi      r3, r4, 32
78          blr
79EPILOGUE()
80