URL
https://opencores.org/ocsvn/forwardcom/forwardcom/trunk
Subversion Repositories forwardcom
[/] [forwardcom/] [libraries/] [divide_int_light.as] - Rev 85
Go to most recent revision | Compare with Previous | Blame | View Log
/**************************** divide_int_light.as **************************
* Author: Agner Fog
* date created: 2021-05-26
* Last modified: 2021-05-26
* Version: 1.11
* Project: ForwardCom library libc_light.li
* Description: divide_int: divide two 32-bit signed integers
* This function is for small CPUs with limited capabilities and no division
* instruction. It returns the quotient and the remainder.
* Returns INT_MIN if dividing by zero.
*
* Copyright 2021 GNU General Public License http://www.gnu.org/licenses
*****************************************************************************/
code section execute
_divide_int function public reguse=3,0
// divide two 32 bit signed integers. return quotient and remainder
// input r0: dividend
// input r1: divisor
// output r0: r0 / r1
// output r1: r0 % r1
if (int32 r1 == 0) {jump divide_error} // check for division by 0
// save r2 - r5
int64 sp -= 4*8
int64 [sp+0x00] = r2
int64 [sp+0x08] = r3
int64 [sp+0x10] = r4
int64 [sp+0x18] = r5
// get signs
int32 r2 = r0 < 0 // sign of r0
int32 r0 = -r0, mask = r2 // abs(r0)
int32 r3 = r1 < 0 // sign of r1
int32 r1 = -r1, mask = r3 // abs(r1)
int r3 ^= r2 // sign of result
int r2 = 0 // quotient calcuated in r2
int r4 = bitscan(r0, 1) // number of significant bits in dividend
int r5 = bitscan(r1, 1) // number of significant bits in divisor
if (int r4 >= r5) {
// division loop
int r4 -= r5 // approximate number of significant bits in quotient
int r1 <<= r4 // shift left divisor
do { // loop r4 + 1 times
uint32 r5 = r0 >= r1 // one bit of quotient
int32 r0 -= r1, mask = r5 // subtract if bigger
int32 r2 <<= 1 // shift left quotient
int32 r2 |= r5 // add new bit
uint32 r1 >>= 1 // shift right divisor
int r4-- // loop counter
} while (int r4 >= 0)
}
// quotient = r2, remainder = r0
int32 r1 = r3 ? -r0 : r0 // apply sign to remainder
int32 r0 = r3 ? -r2 : r2 // apply sign to quotient
// restore r2 - r5
int64 r2 = [sp+0x00]
int64 r3 = [sp+0x08]
int64 r4 = [sp+0x10]
int64 r5 = [sp+0x18]
int64 sp += 4*8
return
divide_error: // division by zero
int32 r0 = 0x80000000 // return INT_MIN
int32 r1 = 0
return
code end
Go to most recent revision | Compare with Previous | Blame | View Log