1/* hppa1.1 addmul_1 -- Multiply a limb vector with a limb and add 2 * the result to a second limb vector. 3 * 4 * Copyright (C) 1992, 1993, 1994, 1998, 5 * 2001, 2002 Free Software Foundation, Inc. 6 * 7 * This file is part of Libgcrypt. 8 * 9 * Libgcrypt is free software; you can redistribute it and/or modify 10 * it under the terms of the GNU Lesser General Public License as 11 * published by the Free Software Foundation; either version 2.1 of 12 * the License, or (at your option) any later version. 13 * 14 * Libgcrypt 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 Lesser General Public License for more details. 18 * 19 * You should have received a copy of the GNU Lesser General Public 20 * License along with this program; if not, write to the Free Software 21 * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA 22 * 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 */ 29 30 31/******************* 32 * mpi_limb_t 33 * _gcry_mpih_addmul_1( mpi_ptr_t res_ptr, (r26) 34 * mpi_ptr_t s1_ptr, (r25) 35 * mpi_size_t s1_size, (r24) 36 * mpi_limb_t s2_limb) (r23) 37 * 38 * This runs at 11 cycles/limb on a PA7000. With the used instructions, it 39 * can not become faster due to data cache contention after a store. On the 40 * PA7100 it runs at 10 cycles/limb, and that can not be improved either, 41 * since only the xmpyu does not need the integer pipeline, so the only 42 * dual-issue we will get are addc+xmpyu. Unrolling could gain a cycle/limb 43 * on the PA7100. 44 * 45 * There are some ideas described in mul1.S that applies to this code too. 46 */ 47 48 .level 1.1 49 50 .code 51 .export _gcry_mpih_addmul_1 52 .label _gcry_mpih_addmul_1 53 .proc 54 .callinfo frame=64,no_calls 55 .entry 56 57 ldo 64(%r30),%r30 58 fldws,ma 4(%r25),%fr5 59 stw %r23,-16(%r30) ; move s2_limb ... 60 addib,= -1,%r24,L$just_one_limb 61 fldws -16(%r30),%fr4 ; ... into fr4 62 add %r0,%r0,%r0 ; clear carry 63 xmpyu %fr4,%fr5,%fr6 64 fldws,ma 4(%r25),%fr7 65 fstds %fr6,-16(%r30) 66 xmpyu %fr4,%fr7,%fr8 67 ldw -12(%r30),%r19 ; least significant limb in product 68 ldw -16(%r30),%r28 69 70 fstds %fr8,-16(%r30) 71 addib,= -1,%r24,L$end 72 ldw -12(%r30),%r1 73 74; Main loop 75 .label L$loop 76 ldws 0(%r26),%r29 77 fldws,ma 4(%r25),%fr5 78 add %r29,%r19,%r19 79 stws,ma %r19,4(%r26) 80 addc %r28,%r1,%r19 81 xmpyu %fr4,%fr5,%fr6 82 ldw -16(%r30),%r28 83 fstds %fr6,-16(%r30) 84 addc %r0,%r28,%r28 85 addib,<> -1,%r24,L$loop 86 ldw -12(%r30),%r1 87 88 .label L$end 89 ldw 0(%r26),%r29 90 add %r29,%r19,%r19 91 stws,ma %r19,4(%r26) 92 addc %r28,%r1,%r19 93 ldw -16(%r30),%r28 94 ldws 0(%r26),%r29 95 addc %r0,%r28,%r28 96 add %r29,%r19,%r19 97 stws,ma %r19,4(%r26) 98 addc %r0,%r28,%r28 99 bv 0(%r2) 100 ldo -64(%r30),%r30 101 102 .label L$just_one_limb 103 xmpyu %fr4,%fr5,%fr6 104 ldw 0(%r26),%r29 105 fstds %fr6,-16(%r30) 106 ldw -12(%r30),%r1 107 ldw -16(%r30),%r28 108 add %r29,%r1,%r19 109 stw %r19,0(%r26) 110 addc %r0,%r28,%r28 111 bv 0(%r2) 112 ldo -64(%r30),%r30 113 114 .exit 115 .procend 116 117 118