| 1 |
###################################- |
| 2 |
# |
| 3 |
# Copyright (C) 2009-2025 Free Software Foundation, Inc. |
| 4 |
# |
| 5 |
# Contributed by Michael Eager <[email protected]>. |
| 6 |
# |
| 7 |
# This file is free software; you can redistribute it and/or modify it |
| 8 |
# under the terms of the GNU General Public License as published by the |
| 9 |
# Free Software Foundation; either version 3, or (at your option) any |
| 10 |
# later version. |
| 11 |
# |
| 12 |
# GCC is distributed in the hope that it will be useful, but WITHOUT |
| 13 |
# ANY WARRANTY; without even the implied warranty of MERCHANTABILITY |
| 14 |
# or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public |
| 15 |
# License for more details. |
| 16 |
# |
| 17 |
# Under Section 7 of GPL version 3, you are granted additional |
| 18 |
# permissions described in the GCC Runtime Library Exception, version |
| 19 |
# 3.1, as published by the Free Software Foundation. |
| 20 |
# |
| 21 |
# You should have received a copy of the GNU General Public License and |
| 22 |
# a copy of the GCC Runtime Library Exception along with this program; |
| 23 |
# see the files COPYING3 and COPYING.RUNTIME respectively. If not, see |
| 24 |
# <http://www.gnu.org/licenses/>. |
| 25 |
# |
| 26 |
# muldi3_hard.S |
| 27 |
# |
| 28 |
# Multiply operation for 64 bit integers, for devices with hard multiply |
| 29 |
# Input : Operand1[H] in Reg r5 |
| 30 |
# Operand1[L] in Reg r6 |
| 31 |
# Operand2[H] in Reg r7 |
| 32 |
# Operand2[L] in Reg r8 |
| 33 |
# Output: Result[H] in Reg r3 |
| 34 |
# Result[L] in Reg r4 |
| 35 |
# |
| 36 |
# Explaination: |
| 37 |
# |
| 38 |
# Both the input numbers are divided into 16 bit number as follows |
| 39 |
# op1 = A B C D |
| 40 |
# op2 = E F G H |
| 41 |
# result = D * H |
| 42 |
# + (C * H + D * G) << 16 |
| 43 |
# + (B * H + C * G + D * F) << 32 |
| 44 |
# + (A * H + B * G + C * F + D * E) << 48 |
| 45 |
# |
| 46 |
# Only 64 bits of the output are considered |
| 47 |
# |
| 48 |
####################################### |
| 49 |
|
| 50 |
/* An executable stack is *not* required for these functions. */ |
| 51 |
#ifdef __linux__ |
| 52 |
.section .note.GNU-stack,"",%progbits |
| 53 |
.previous |
| 54 |
#endif |
| 55 |
|
| 56 |
.globl muldi3_hardproc |
| 57 |
.ent muldi3_hardproc |
| 58 |
muldi3_hardproc: |
| 59 |
addi r1,r1,-40 |
| 60 |
|
| 61 |
# Save the input operands on the caller's stack |
| 62 |
swi r5,r1,44 |
| 63 |
swi r6,r1,48 |
| 64 |
swi r7,r1,52 |
| 65 |
swi r8,r1,56 |
| 66 |
|
| 67 |
# Store all the callee saved registers |
| 68 |
sw r20,r1,r0 |
| 69 |
swi r21,r1,4 |
| 70 |
swi r22,r1,8 |
| 71 |
swi r23,r1,12 |
| 72 |
swi r24,r1,16 |
| 73 |
swi r25,r1,20 |
| 74 |
swi r26,r1,24 |
| 75 |
swi r27,r1,28 |
| 76 |
|
| 77 |
# Load all the 16 bit values for A through H |
| 78 |
lhui r20,r1,44 # A |
| 79 |
lhui r21,r1,46 # B |
| 80 |
lhui r22,r1,48 # C |
| 81 |
lhui r23,r1,50 # D |
| 82 |
lhui r24,r1,52 # E |
| 83 |
lhui r25,r1,54 # F |
| 84 |
lhui r26,r1,56 # G |
| 85 |
lhui r27,r1,58 # H |
| 86 |
|
| 87 |
# D * H ==> LSB of the result on stack ==> Store1 |
| 88 |
mul r9,r23,r27 |
| 89 |
swi r9,r1,36 # Pos2 and Pos3 |
| 90 |
|
| 91 |
# Hi (Store1) + C * H + D * G ==> Store2 ==> Pos1 and Pos2 |
| 92 |
# Store the carry generated in position 2 for Pos 3 |
| 93 |
lhui r11,r1,36 # Pos2 |
| 94 |
mul r9,r22,r27 # C * H |
| 95 |
mul r10,r23,r26 # D * G |
| 96 |
add r9,r9,r10 |
| 97 |
addc r12,r0,r0 |
| 98 |
add r9,r9,r11 |
| 99 |
addc r12,r12,r0 # Store the Carry |
| 100 |
shi r9,r1,36 # Store Pos2 |
| 101 |
swi r9,r1,32 |
| 102 |
lhui r11,r1,32 |
| 103 |
shi r11,r1,34 # Store Pos1 |
| 104 |
|
| 105 |
# Hi (Store2) + B * H + C * G + D * F ==> Store3 ==> Pos0 and Pos1 |
| 106 |
mul r9,r21,r27 # B * H |
| 107 |
mul r10,r22,r26 # C * G |
| 108 |
mul r7,r23,r25 # D * F |
| 109 |
add r9,r9,r11 |
| 110 |
add r9,r9,r10 |
| 111 |
add r9,r9,r7 |
| 112 |
swi r9,r1,32 # Pos0 and Pos1 |
| 113 |
|
| 114 |
# Hi (Store3) + A * H + B * G + C * F + D * E ==> Store3 ==> Pos0 |
| 115 |
lhui r11,r1,32 # Pos0 |
| 116 |
mul r9,r20,r27 # A * H |
| 117 |
mul r10,r21,r26 # B * G |
| 118 |
mul r7,r22,r25 # C * F |
| 119 |
mul r8,r23,r24 # D * E |
| 120 |
add r9,r9,r11 |
| 121 |
add r9,r9,r10 |
| 122 |
add r9,r9,r7 |
| 123 |
add r9,r9,r8 |
| 124 |
sext16 r9,r9 # Sign extend the MSB |
| 125 |
shi r9,r1,32 |
| 126 |
|
| 127 |
# Move results to r3 and r4 |
| 128 |
lhui r3,r1,32 |
| 129 |
add r3,r3,r12 |
| 130 |
shi r3,r1,32 |
| 131 |
lwi r3,r1,32 # Hi Part |
| 132 |
lwi r4,r1,36 # Lo Part |
| 133 |
|
| 134 |
# Restore Callee saved registers |
| 135 |
lw r20,r1,r0 |
| 136 |
lwi r21,r1,4 |
| 137 |
lwi r22,r1,8 |
| 138 |
lwi r23,r1,12 |
| 139 |
lwi r24,r1,16 |
| 140 |
lwi r25,r1,20 |
| 141 |
lwi r26,r1,24 |
| 142 |
lwi r27,r1,28 |
| 143 |
|
| 144 |
# Restore Frame and return |
| 145 |
rtsd r15,8 |
| 146 |
addi r1,r1,40 |
| 147 |
|
| 148 |
.end muldi3_hardproc |
| 149 |
|
| 150 |
|