1
/* hppa1.1 submul_1 -- Multiply a limb vector with a limb and add
2
* the result to a second limb vector.
4
* Copyright (C) 1992, 1993, 1994, 1998,
5
* 2001 Free Software Foundation, Inc.
7
* This file is part of GnuPG.
9
* GnuPG is free software; you can redistribute it and/or modify
10
* it under the terms of the GNU General Public License as published by
11
* the Free Software Foundation; either version 2 of the License, or
12
* (at your option) any later version.
14
* GnuPG is distributed in the hope that it will be useful,
15
* but WITHOUT ANY WARRANTY; without even the implied warranty of
16
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
17
* GNU General Public License for more details.
19
* You should have received a copy of the GNU General Public License
20
* along with this program; if not, write to the Free Software
21
* Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA
23
* Note: This code is heavily based on the GNU MP Library.
24
* Actually it's the same code with only minor changes in the
25
* way the data is stored; this is to support the abstraction
26
* of an optional secure memory allocation which may be used
27
* to avoid revealing of sensitive data due to paging etc.
28
* The GNU MP Library itself is published under the LGPL;
29
* however I decided to publish this code under the plain GPL.
35
* mpihelp_submul_1( mpi_ptr_t res_ptr, (r26)
36
* mpi_ptr_t s1_ptr, (r25)
37
* mpi_size_t s1_size, (r24)
38
* mpi_limb_t s2_limb) (r23)
41
* This runs at 12 cycles/limb on a PA7000. With the used instructions, it
42
* can not become faster due to data cache contention after a store. On the
43
* PA7100 it runs at 11 cycles/limb, and that can not be improved either,
44
* since only the xmpyu does not need the integer pipeline, so the only
45
* dual-issue we will get are addc+xmpyu. Unrolling could gain a cycle/limb
48
* There are some ideas described in mul1.S that applies to this code too.
50
* It seems possible to make this run as fast as addmul_1, if we use
51
* sub,>>= %r29,%r19,%r22
53
* but that requires reworking the hairy software pipeline...
58
.export mpihelp_submul_1
59
.label mpihelp_submul_1
61
.callinfo frame=64,no_calls
66
stw %r23,-16(%r30) ; move s2_limb ...
67
addib,= -1,%r24,L$just_one_limb
68
fldws -16(%r30),%fr4 ; ... into fr4
69
add %r0,%r0,%r0 ; clear carry
74
ldw -12(%r30),%r19 ; least significant limb in product
93
addib,<> -1,%r24,L$loop
112
.label L$just_one_limb