208 lines
5.1 KiB
C++
208 lines
5.1 KiB
C++
|
/* armsupp.c -- ARMulator support code: ARM6 Instruction Emulator.
|
||
|
Copyright (C) 1994 Advanced RISC Machines Ltd.
|
||
|
|
||
|
This program is free software; you can redistribute it and/or modify
|
||
|
it under the terms of the GNU General Public License as published by
|
||
|
the Free Software Foundation; either version 2 of the License, or
|
||
|
(at your option) any later version.
|
||
|
|
||
|
This program is distributed in the hope that it will be useful,
|
||
|
but WITHOUT ANY WARRANTY; without even the implied warranty of
|
||
|
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
||
|
GNU General Public License for more details.
|
||
|
|
||
|
You should have received a copy of the GNU General Public License
|
||
|
along with this program; if not, write to the Free Software
|
||
|
Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */
|
||
|
|
||
|
#include "common/logging/log.h"
|
||
|
|
||
|
#include "tests/skyeye_interpreter/skyeye_common/arm_regformat.h"
|
||
|
#include "tests/skyeye_interpreter/skyeye_common/armstate.h"
|
||
|
#include "tests/skyeye_interpreter/skyeye_common/armsupp.h"
|
||
|
|
||
|
// Unsigned sum of absolute difference
|
||
|
u8 ARMul_UnsignedAbsoluteDifference(u8 left, u8 right)
|
||
|
{
|
||
|
if (left > right)
|
||
|
return left - right;
|
||
|
|
||
|
return right - left;
|
||
|
}
|
||
|
|
||
|
// Add with carry, indicates if a carry-out or signed overflow occurred.
|
||
|
u32 AddWithCarry(u32 left, u32 right, u32 carry_in, bool* carry_out_occurred, bool* overflow_occurred)
|
||
|
{
|
||
|
u64 unsigned_sum = (u64)left + (u64)right + (u64)carry_in;
|
||
|
s64 signed_sum = (s64)(s32)left + (s64)(s32)right + (s64)carry_in;
|
||
|
u64 result = (unsigned_sum & 0xFFFFFFFF);
|
||
|
|
||
|
if (carry_out_occurred)
|
||
|
*carry_out_occurred = (result != unsigned_sum);
|
||
|
|
||
|
if (overflow_occurred)
|
||
|
*overflow_occurred = ((s64)(s32)result != signed_sum);
|
||
|
|
||
|
return (u32)result;
|
||
|
}
|
||
|
|
||
|
// Compute whether an addition of A and B, giving RESULT, overflowed.
|
||
|
bool AddOverflow(u32 a, u32 b, u32 result)
|
||
|
{
|
||
|
return ((NEG(a) && NEG(b) && POS(result)) ||
|
||
|
(POS(a) && POS(b) && NEG(result)));
|
||
|
}
|
||
|
|
||
|
// Compute whether a subtraction of A and B, giving RESULT, overflowed.
|
||
|
bool SubOverflow(u32 a, u32 b, u32 result)
|
||
|
{
|
||
|
return ((NEG(a) && POS(b) && POS(result)) ||
|
||
|
(POS(a) && NEG(b) && NEG(result)));
|
||
|
}
|
||
|
|
||
|
// Returns true if the Q flag should be set as a result of overflow.
|
||
|
bool ARMul_AddOverflowQ(u32 a, u32 b)
|
||
|
{
|
||
|
u32 result = a + b;
|
||
|
if (((result ^ a) & (u32)0x80000000) && ((a ^ b) & (u32)0x80000000) == 0)
|
||
|
return true;
|
||
|
|
||
|
return false;
|
||
|
}
|
||
|
|
||
|
// 8-bit signed saturated addition
|
||
|
u8 ARMul_SignedSaturatedAdd8(u8 left, u8 right)
|
||
|
{
|
||
|
u8 result = left + right;
|
||
|
|
||
|
if (((result ^ left) & 0x80) && ((left ^ right) & 0x80) == 0) {
|
||
|
if (left & 0x80)
|
||
|
result = 0x80;
|
||
|
else
|
||
|
result = 0x7F;
|
||
|
}
|
||
|
|
||
|
return result;
|
||
|
}
|
||
|
|
||
|
// 8-bit signed saturated subtraction
|
||
|
u8 ARMul_SignedSaturatedSub8(u8 left, u8 right)
|
||
|
{
|
||
|
u8 result = left - right;
|
||
|
|
||
|
if (((result ^ left) & 0x80) && ((left ^ right) & 0x80) != 0) {
|
||
|
if (left & 0x80)
|
||
|
result = 0x80;
|
||
|
else
|
||
|
result = 0x7F;
|
||
|
}
|
||
|
|
||
|
return result;
|
||
|
}
|
||
|
|
||
|
// 16-bit signed saturated addition
|
||
|
u16 ARMul_SignedSaturatedAdd16(u16 left, u16 right)
|
||
|
{
|
||
|
u16 result = left + right;
|
||
|
|
||
|
if (((result ^ left) & 0x8000) && ((left ^ right) & 0x8000) == 0) {
|
||
|
if (left & 0x8000)
|
||
|
result = 0x8000;
|
||
|
else
|
||
|
result = 0x7FFF;
|
||
|
}
|
||
|
|
||
|
return result;
|
||
|
}
|
||
|
|
||
|
// 16-bit signed saturated subtraction
|
||
|
u16 ARMul_SignedSaturatedSub16(u16 left, u16 right)
|
||
|
{
|
||
|
u16 result = left - right;
|
||
|
|
||
|
if (((result ^ left) & 0x8000) && ((left ^ right) & 0x8000) != 0) {
|
||
|
if (left & 0x8000)
|
||
|
result = 0x8000;
|
||
|
else
|
||
|
result = 0x7FFF;
|
||
|
}
|
||
|
|
||
|
return result;
|
||
|
}
|
||
|
|
||
|
// 8-bit unsigned saturated addition
|
||
|
u8 ARMul_UnsignedSaturatedAdd8(u8 left, u8 right)
|
||
|
{
|
||
|
u8 result = left + right;
|
||
|
|
||
|
if (result < left)
|
||
|
result = 0xFF;
|
||
|
|
||
|
return result;
|
||
|
}
|
||
|
|
||
|
// 16-bit unsigned saturated addition
|
||
|
u16 ARMul_UnsignedSaturatedAdd16(u16 left, u16 right)
|
||
|
{
|
||
|
u16 result = left + right;
|
||
|
|
||
|
if (result < left)
|
||
|
result = 0xFFFF;
|
||
|
|
||
|
return result;
|
||
|
}
|
||
|
|
||
|
// 8-bit unsigned saturated subtraction
|
||
|
u8 ARMul_UnsignedSaturatedSub8(u8 left, u8 right)
|
||
|
{
|
||
|
if (left <= right)
|
||
|
return 0;
|
||
|
|
||
|
return left - right;
|
||
|
}
|
||
|
|
||
|
// 16-bit unsigned saturated subtraction
|
||
|
u16 ARMul_UnsignedSaturatedSub16(u16 left, u16 right)
|
||
|
{
|
||
|
if (left <= right)
|
||
|
return 0;
|
||
|
|
||
|
return left - right;
|
||
|
}
|
||
|
|
||
|
// Signed saturation.
|
||
|
u32 ARMul_SignedSatQ(s32 value, u8 shift, bool* saturation_occurred)
|
||
|
{
|
||
|
const u32 max = (1 << shift) - 1;
|
||
|
const s32 top = (value >> shift);
|
||
|
|
||
|
if (top > 0) {
|
||
|
*saturation_occurred = true;
|
||
|
return max;
|
||
|
}
|
||
|
else if (top < -1) {
|
||
|
*saturation_occurred = true;
|
||
|
return ~max;
|
||
|
}
|
||
|
|
||
|
*saturation_occurred = false;
|
||
|
return (u32)value;
|
||
|
}
|
||
|
|
||
|
// Unsigned saturation
|
||
|
u32 ARMul_UnsignedSatQ(s32 value, u8 shift, bool* saturation_occurred)
|
||
|
{
|
||
|
const u32 max = (1 << shift) - 1;
|
||
|
|
||
|
if (value < 0) {
|
||
|
*saturation_occurred = true;
|
||
|
return 0;
|
||
|
} else if ((u32)value > max) {
|
||
|
*saturation_occurred = true;
|
||
|
return max;
|
||
|
}
|
||
|
|
||
|
*saturation_occurred = false;
|
||
|
return (u32)value;
|
||
|
}
|