mirror of https://github.com/tasks/tasks
You cannot select more than 25 topics
Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
1172 lines
23 KiB
C
1172 lines
23 KiB
C
/*
|
|
** Copyright 2003-2010, VisualOn, Inc.
|
|
**
|
|
** Licensed under the Apache License, Version 2.0 (the "License");
|
|
** you may not use this file except in compliance with the License.
|
|
** You may obtain a copy of the License at
|
|
**
|
|
** http://www.apache.org/licenses/LICENSE-2.0
|
|
**
|
|
** Unless required by applicable law or agreed to in writing, software
|
|
** distributed under the License is distributed on an "AS IS" BASIS,
|
|
** WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
|
** See the License for the specific language governing permissions and
|
|
** limitations under the License.
|
|
*/
|
|
/*******************************************************************************
|
|
File: basicop2.h
|
|
|
|
Content: Constants , Globals and Basic arithmetic operators.
|
|
|
|
*******************************************************************************/
|
|
|
|
#ifndef __BASIC_OP_H
|
|
#define __BASIC_OP_H
|
|
|
|
#include "typedef.h"
|
|
|
|
#define MAX_32 (Word32)0x7fffffffL
|
|
#define MIN_32 (Word32)0x80000000L
|
|
|
|
#define MAX_16 (Word16)0x7fff
|
|
#define MIN_16 (Word16)0x8000
|
|
#define ABS(a) ((a) >= 0) ? (a) : (-(a))
|
|
|
|
/* Short abs, 1 */
|
|
#define abs_s(x) (((x) != MIN_16) ? (((x) >= 0) ? (x) : (-(x))) : MAX_16)
|
|
|
|
/* 16 bit var1 -> MSB, 2 */
|
|
#define L_deposit_h(x) (((Word32)(x)) << 16)
|
|
|
|
|
|
/* 16 bit var1 -> LSB, 2 */
|
|
#define L_deposit_l(x) ((Word32)(x))
|
|
|
|
|
|
/* Long abs, 3 */
|
|
#define L_abs(x) (((x) != MIN_32) ? (((x) >= 0) ? (x) : (-(x))) : MAX_32)
|
|
|
|
|
|
/* Short negate, 1 */
|
|
#define negate(var1) (((var1) == MIN_16) ? MAX_16 : (-(var1)))
|
|
|
|
|
|
/* Long negate, 2 */
|
|
#define L_negate(L_var1) (((L_var1) == (MIN_32)) ? (MAX_32) : (-(L_var1)))
|
|
|
|
|
|
#define MULHIGH(A,B) (int)(((Word64)(A)*(Word64)(B)) >> 32)
|
|
#define fixmul(a, b) (int)((((Word64)(a)*(Word64)(b)) >> 32) << 1)
|
|
|
|
|
|
#if (SATRUATE_IS_INLINE)
|
|
__inline Word32 saturate(Word32 L_var1);
|
|
#else
|
|
Word16 saturate(Word32 L_var1);
|
|
#endif
|
|
|
|
/* Short shift left, 1 */
|
|
#if (SHL_IS_INLINE)
|
|
__inline Word32 shl (Word32 var1, Word32 var2);
|
|
#else
|
|
Word16 shl (Word16 var1, Word16 var2);
|
|
#endif
|
|
|
|
/* Short shift right, 1 */
|
|
#if (SHR_IS_INLINE)
|
|
__inline Word32 shr (Word32 var1, Word32 var2);
|
|
#else
|
|
Word16 shr (Word16 var1, Word16 var2);
|
|
#endif
|
|
|
|
#if (L_MULT_IS_INLINE)
|
|
__inline Word32 L_mult(Word32 var1, Word32 var2);
|
|
#else
|
|
Word32 L_mult(Word16 var1, Word16 var2);
|
|
#endif
|
|
|
|
/* Msu, 1 */
|
|
#if (L_MSU_IS_INLINE)
|
|
__inline Word32 L_msu (Word32 L_var3, Word32 var1, Word32 var2);
|
|
#else
|
|
Word32 L_msu (Word32 L_var3, Word16 var1, Word16 var2);
|
|
#endif
|
|
|
|
/* Long sub, 2 */
|
|
#if (L_SUB_IS_INLINE)
|
|
__inline Word32 L_sub(Word32 L_var1, Word32 L_var2);
|
|
#else
|
|
Word32 L_sub(Word32 L_var1, Word32 L_var2);
|
|
#endif
|
|
|
|
/* Long shift left, 2 */
|
|
#if (L_SHL_IS_INLINE)
|
|
__inline Word32 L_shl (Word32 L_var1, Word32 var2);
|
|
#else
|
|
Word32 L_shl (Word32 L_var1, Word16 var2);
|
|
#endif
|
|
|
|
/* Long shift right, 2*/
|
|
#if (L_SHR_IS_INLINE)
|
|
__inline Word32 L_shr (Word32 L_var1, Word32 var2);
|
|
#else
|
|
Word32 L_shr (Word32 L_var1, Word16 var2);
|
|
#endif
|
|
|
|
/* Short add, 1 */
|
|
#if (ADD_IS_INLINE)
|
|
__inline Word32 add (Word32 var1, Word32 var2);
|
|
#else
|
|
Word16 add (Word16 var1, Word16 var2);
|
|
#endif
|
|
|
|
/* Short sub, 1 */
|
|
#if (SUB_IS_INLINE)
|
|
__inline Word32 sub(Word32 var1, Word32 var2);
|
|
#else
|
|
Word16 sub(Word16 var1, Word16 var2);
|
|
#endif
|
|
|
|
/* Short division, 18 */
|
|
#if (DIV_S_IS_INLINE)
|
|
__inline Word32 div_s (Word32 var1, Word32 var2);
|
|
#else
|
|
Word16 div_s (Word16 var1, Word16 var2);
|
|
#endif
|
|
|
|
/* Short mult, 1 */
|
|
#if (MULT_IS_INLINE)
|
|
__inline Word32 mult (Word32 var1, Word32 var2);
|
|
#else
|
|
Word16 mult (Word16 var1, Word16 var2);
|
|
#endif
|
|
|
|
/* Short norm, 15 */
|
|
#if (NORM_S_IS_INLINE)
|
|
__inline Word32 norm_s (Word32 var1);
|
|
#else
|
|
Word16 norm_s (Word16 var1);
|
|
#endif
|
|
|
|
/* Long norm, 30 */
|
|
#if (NORM_L_IS_INLINE)
|
|
__inline Word32 norm_l (Word32 L_var1);
|
|
#else
|
|
Word16 norm_l (Word32 L_var1);
|
|
#endif
|
|
|
|
/* Round, 1 */
|
|
#if (ROUND_IS_INLINE)
|
|
__inline Word32 round16(Word32 L_var1);
|
|
#else
|
|
Word16 round16(Word32 L_var1);
|
|
#endif
|
|
|
|
/* Mac, 1 */
|
|
#if (L_MAC_IS_INLINE)
|
|
__inline Word32 L_mac (Word32 L_var3, Word32 var1, Word32 var2);
|
|
#else
|
|
Word32 L_mac (Word32 L_var3, Word16 var1, Word16 var2);
|
|
#endif
|
|
|
|
#if (L_ADD_IS_INLINE)
|
|
__inline Word32 L_add (Word32 L_var1, Word32 L_var2);
|
|
#else
|
|
Word32 L_add (Word32 L_var1, Word32 L_var2);
|
|
#endif
|
|
|
|
/* Extract high, 1 */
|
|
#if (EXTRACT_H_IS_INLINE)
|
|
__inline Word32 extract_h (Word32 L_var1);
|
|
#else
|
|
Word16 extract_h (Word32 L_var1);
|
|
#endif
|
|
|
|
/* Extract low, 1 */
|
|
#if (EXTRACT_L_IS_INLINE)
|
|
__inline Word32 extract_l(Word32 L_var1);
|
|
#else
|
|
Word16 extract_l(Word32 L_var1);
|
|
#endif
|
|
|
|
/* Mult with round, 2 */
|
|
#if (MULT_R_IS_INLINE)
|
|
__inline Word32 mult_r(Word32 var1, Word32 var2);
|
|
#else
|
|
Word16 mult_r(Word16 var1, Word16 var2);
|
|
#endif
|
|
|
|
/* Shift right with round, 2 */
|
|
#if (SHR_R_IS_INLINE)
|
|
__inline Word32 shr_r (Word32 var1, Word32 var2);
|
|
#else
|
|
Word16 shr_r (Word16 var1, Word16 var2);
|
|
#endif
|
|
|
|
/* Mac with rounding,2 */
|
|
#if (MAC_R_IS_INLINE)
|
|
__inline Word32 mac_r (Word32 L_var3, Word32 var1, Word32 var2);
|
|
#else
|
|
Word16 mac_r (Word32 L_var3, Word16 var1, Word16 var2);
|
|
#endif
|
|
|
|
/* Msu with rounding,2 */
|
|
#if (MSU_R_IS_INLINE)
|
|
__inline Word32 msu_r (Word32 L_var3, Word32 var1, Word32 var2);
|
|
#else
|
|
Word16 msu_r (Word32 L_var3, Word16 var1, Word16 var2);
|
|
#endif
|
|
|
|
/* Long shift right with round, 3 */
|
|
#if (L_SHR_R_IS_INLINE)
|
|
__inline Word32 L_shr_r (Word32 L_var1, Word32 var2);
|
|
#else
|
|
Word32 L_shr_r (Word32 L_var1, Word16 var2);
|
|
#endif
|
|
|
|
#if ARMV4_INASM
|
|
__inline Word32 ASM_L_shr(Word32 L_var1, Word32 var2)
|
|
{
|
|
return L_var1 >> var2;
|
|
}
|
|
|
|
__inline Word32 ASM_L_shl(Word32 L_var1, Word32 var2)
|
|
{
|
|
Word32 result;
|
|
asm (
|
|
"MOV %[result], %[L_var1], ASL %[var2] \n"
|
|
"TEQ %[L_var1], %[result], ASR %[var2]\n"
|
|
"EORNE %[result], %[mask], %[L_var1], ASR #31\n"
|
|
:[result]"=&r"(result)
|
|
:[L_var1]"r"(L_var1), [var2]"r"(var2), [mask]"r"(0x7fffffff)
|
|
);
|
|
return result;
|
|
}
|
|
|
|
__inline Word32 ASM_shr(Word32 L_var1, Word32 var2)
|
|
{
|
|
Word32 result;
|
|
asm (
|
|
"CMP %[var2], #15\n"
|
|
"MOVLT %[result], %[L_var1], ASR %[var2]\n"
|
|
"MOVGE %[result], %[L_var1], ASR #15\n"
|
|
:[result]"=r"(result)
|
|
:[L_var1]"r"(L_var1), [var2]"r"(var2)
|
|
);
|
|
return result;
|
|
}
|
|
|
|
__inline Word32 ASM_shl(Word32 L_var1, Word32 var2)
|
|
{
|
|
#if ARMV6_SAT
|
|
Word32 result;
|
|
asm (
|
|
"CMP %[var2], #16\n"
|
|
"MOVLT %[result], %[L_var1], ASL %[var2]\n"
|
|
"MOVGE %[result], %[L_var1], ASL #16\n"
|
|
"SSAT %[result], #16, %[result]\n"
|
|
:[result]"=r"(result)
|
|
:[L_var1]"r"(L_var1), [var2]"r"(var2)
|
|
);
|
|
return result;
|
|
#else
|
|
Word32 result;
|
|
Word32 tmp;
|
|
asm (
|
|
"CMP %[var2], #16\n"
|
|
"MOVLT %[result], %[L_var1], ASL %[var2]\n"
|
|
"MOVGE %[result], %[L_var1], ASL #16\n"
|
|
"MOV %[tmp], %[result], ASR #15\n"
|
|
"TEQ %[tmp], %[result], ASR #31 \n"
|
|
"EORNE %[result], %[mask], %[result],ASR #31"
|
|
:[result]"=&r"(result), [tmp]"=&r"(tmp)
|
|
:[L_var1]"r"(L_var1), [var2]"r"(var2), [mask]"r"(0x7fff)
|
|
);
|
|
return result;
|
|
#endif
|
|
}
|
|
#endif
|
|
|
|
/*___________________________________________________________________________
|
|
| |
|
|
| definitions for inline basic arithmetic operators |
|
|
|___________________________________________________________________________|
|
|
*/
|
|
#if (SATRUATE_IS_INLINE)
|
|
__inline Word32 saturate(Word32 L_var1)
|
|
{
|
|
#if ARMV6_SAT
|
|
Word32 result;
|
|
asm (
|
|
"SSAT %[result], #16, %[L_var1]"
|
|
: [result]"=r"(result)
|
|
: [L_var1]"r"(L_var1)
|
|
);
|
|
return result;
|
|
#elif ARMV5TE_SAT
|
|
Word32 result;
|
|
Word32 tmp;
|
|
asm volatile (
|
|
"MOV %[tmp], %[L_var1],ASR#15\n"
|
|
"TEQ %[tmp], %[L_var1],ASR#31\n"
|
|
"EORNE %[result], %[mask],%[L_var1],ASR#31\n"
|
|
"MOVEQ %[result], %[L_var1]\n"
|
|
:[result]"=&r"(result), [tmp]"=&r"(tmp)
|
|
:[L_var1]"r"(L_var1), [mask]"r"(0x7fff)
|
|
);
|
|
|
|
return result;
|
|
#else
|
|
Word32 var_out;
|
|
|
|
//var_out = (L_var1 > (Word32)0X00007fffL) ? (MAX_16) : ((L_var1 < (Word32)0xffff8000L) ? (MIN_16) : ((Word16)L_var1));
|
|
|
|
if (L_var1 > 0X00007fffL)
|
|
{
|
|
var_out = MAX_16;
|
|
}
|
|
else if (L_var1 < (Word32) 0xffff8000L)
|
|
{
|
|
var_out = MIN_16;
|
|
}
|
|
else
|
|
{
|
|
var_out = extract_l(L_var1);
|
|
}
|
|
|
|
return (var_out);
|
|
#endif
|
|
}
|
|
#endif
|
|
|
|
/* Short shift left, 1 */
|
|
#if (SHL_IS_INLINE)
|
|
__inline Word32 shl (Word32 var1, Word32 var2)
|
|
{
|
|
#if ARMV5TE_SHL
|
|
if(var2>=0)
|
|
{
|
|
return ASM_shl( var1, var2);
|
|
}
|
|
else
|
|
{
|
|
return ASM_shr( var1, -var2);
|
|
}
|
|
#else
|
|
Word32 var_out;
|
|
Word32 result;
|
|
|
|
if (var2 < 0)
|
|
{
|
|
var_out = shr (var1, (Word16)-var2);
|
|
}
|
|
else
|
|
{
|
|
result = (Word32) var1 *((Word32) 1 << var2);
|
|
|
|
if ((var2 > 15 && var1 != 0) || (result != (Word32) ((Word16) result)))
|
|
{
|
|
var_out = (Word16)((var1 > 0) ? MAX_16 : MIN_16);
|
|
}
|
|
else
|
|
{
|
|
var_out = extract_l(result);
|
|
}
|
|
}
|
|
return (var_out);
|
|
#endif
|
|
}
|
|
#endif
|
|
|
|
/* Short shift right, 1 */
|
|
#if (SHR_IS_INLINE)
|
|
__inline Word32 shr (Word32 var1, Word32 var2)
|
|
{
|
|
#if ARMV5TE_SHR
|
|
if(var2>=0)
|
|
{
|
|
return ASM_shr( var1, var2);
|
|
}
|
|
else
|
|
{
|
|
return ASM_shl( var1, -var2);
|
|
}
|
|
#else
|
|
Word32 var_out;
|
|
|
|
if (var2 < 0)
|
|
{
|
|
var_out = shl (var1, (Word16)-var2);
|
|
}
|
|
else
|
|
{
|
|
if (var2 >= 15)
|
|
{
|
|
var_out = (Word16)((var1 < 0) ? -1 : 0);
|
|
}
|
|
else
|
|
{
|
|
if (var1 < 0)
|
|
{
|
|
var_out = (Word16)(~((~var1) >> var2));
|
|
}
|
|
else
|
|
{
|
|
var_out = (Word16)(var1 >> var2);
|
|
}
|
|
}
|
|
}
|
|
|
|
return (var_out);
|
|
#endif
|
|
}
|
|
#endif
|
|
|
|
|
|
#if (L_MULT_IS_INLINE)
|
|
__inline Word32 L_mult(Word32 var1, Word32 var2)
|
|
{
|
|
#if ARMV5TE_L_MULT
|
|
Word32 result;
|
|
asm (
|
|
"SMULBB %[result], %[var1], %[var2] \n"
|
|
"QADD %[result], %[result], %[result] \n"
|
|
:[result]"=r"(result)
|
|
:[var1]"r"(var1), [var2]"r"(var2)
|
|
);
|
|
return result;
|
|
#else
|
|
Word32 L_var_out;
|
|
|
|
L_var_out = (Word32) var1 *(Word32) var2;
|
|
|
|
if (L_var_out != (Word32) 0x40000000L)
|
|
{
|
|
L_var_out <<= 1;
|
|
}
|
|
else
|
|
{
|
|
L_var_out = MAX_32;
|
|
}
|
|
return (L_var_out);
|
|
#endif
|
|
}
|
|
#endif
|
|
|
|
#if (L_MSU_IS_INLINE)
|
|
__inline Word32 L_msu (Word32 L_var3, Word32 var1, Word32 var2)
|
|
{
|
|
#if ARMV5TE_L_MSU
|
|
Word32 result;
|
|
asm (
|
|
"SMULBB %[result], %[var1], %[var2] \n"
|
|
"QDSUB %[result], %[L_var3], %[result]\n"
|
|
:[result]"=&r"(result)
|
|
:[L_var3]"r"(L_var3), [var1]"r"(var1), [var2]"r"(var2)
|
|
);
|
|
return result;
|
|
#else
|
|
Word32 L_var_out;
|
|
Word32 L_product;
|
|
|
|
L_product = L_mult(var1, var2);
|
|
L_var_out = L_sub(L_var3, L_product);
|
|
return (L_var_out);
|
|
#endif
|
|
}
|
|
#endif
|
|
|
|
#if (L_SUB_IS_INLINE)
|
|
__inline Word32 L_sub(Word32 L_var1, Word32 L_var2)
|
|
{
|
|
#if ARMV5TE_L_SUB
|
|
Word32 result;
|
|
asm (
|
|
"QSUB %[result], %[L_var1], %[L_var2]\n"
|
|
:[result]"=r"(result)
|
|
:[L_var1]"r"(L_var1), [L_var2]"r"(L_var2)
|
|
);
|
|
return result;
|
|
#else
|
|
Word32 L_var_out;
|
|
|
|
L_var_out = L_var1 - L_var2;
|
|
|
|
if (((L_var1 ^ L_var2) & MIN_32) != 0)
|
|
{
|
|
if ((L_var_out ^ L_var1) & MIN_32)
|
|
{
|
|
L_var_out = (L_var1 < 0L) ? MIN_32 : MAX_32;
|
|
}
|
|
}
|
|
|
|
return (L_var_out);
|
|
#endif
|
|
}
|
|
#endif
|
|
|
|
#if (L_SHL_IS_INLINE)
|
|
__inline Word32 L_shl(Word32 L_var1, Word32 var2)
|
|
{
|
|
#if ARMV5TE_L_SHL
|
|
if(var2>=0)
|
|
{
|
|
return ASM_L_shl( L_var1, var2);
|
|
}
|
|
else
|
|
{
|
|
return ASM_L_shr( L_var1, -var2);
|
|
}
|
|
#else
|
|
Word32 L_var_out = 0L;
|
|
|
|
if (var2 <= 0)
|
|
{
|
|
L_var1 = L_shr(L_var1, (Word16)-var2);
|
|
}
|
|
else
|
|
{
|
|
for (; var2 > 0; var2--)
|
|
{
|
|
if (L_var1 > (Word32) 0X3fffffffL)
|
|
{
|
|
return MAX_32;
|
|
}
|
|
else
|
|
{
|
|
if (L_var1 < (Word32) 0xc0000000L)
|
|
{
|
|
return MIN_32;
|
|
}
|
|
}
|
|
L_var1 <<= 1;
|
|
L_var_out = L_var1;
|
|
}
|
|
}
|
|
return (L_var1);
|
|
#endif
|
|
}
|
|
#endif
|
|
|
|
#if (L_SHR_IS_INLINE)
|
|
__inline Word32 L_shr (Word32 L_var1, Word32 var2)
|
|
{
|
|
#if ARMV5TE_L_SHR
|
|
if(var2>=0)
|
|
{
|
|
return ASM_L_shr( L_var1, var2);
|
|
}
|
|
else
|
|
{
|
|
return ASM_L_shl( L_var1, -var2);
|
|
}
|
|
#else
|
|
Word32 L_var_out;
|
|
|
|
if (var2 < 0)
|
|
{
|
|
L_var_out = L_shl (L_var1, (Word16)-var2);
|
|
}
|
|
else
|
|
{
|
|
if (var2 >= 31)
|
|
{
|
|
L_var_out = (L_var1 < 0L) ? -1 : 0;
|
|
}
|
|
else
|
|
{
|
|
if (L_var1 < 0)
|
|
{
|
|
L_var_out = ~((~L_var1) >> var2);
|
|
}
|
|
else
|
|
{
|
|
L_var_out = L_var1 >> var2;
|
|
}
|
|
}
|
|
}
|
|
return (L_var_out);
|
|
#endif
|
|
}
|
|
#endif
|
|
|
|
/* Short add, 1 */
|
|
#if (ADD_IS_INLINE)
|
|
__inline Word32 add (Word32 var1, Word32 var2)
|
|
{
|
|
#if ARMV5TE_ADD
|
|
Word32 result;
|
|
Word32 tmp;
|
|
asm (
|
|
"ADD %[result], %[var1], %[var2] \n"
|
|
"MOV %[tmp], %[result], ASR #15 \n"
|
|
"TEQ %[tmp], %[result], ASR #31 \n"
|
|
"EORNE %[result], %[mask], %[result], ASR #31"
|
|
:[result]"=&r"(result), [tmp]"=&r"(tmp)
|
|
:[var1]"r"(var1), [var2]"r"(var2), [mask]"r"(0x7fff)
|
|
);
|
|
return result;
|
|
#else
|
|
Word32 var_out;
|
|
Word32 L_sum;
|
|
|
|
L_sum = (Word32) var1 + var2;
|
|
var_out = saturate(L_sum);
|
|
|
|
return (var_out);
|
|
#endif
|
|
}
|
|
#endif
|
|
|
|
/* Short sub, 1 */
|
|
#if (SUB_IS_INLINE)
|
|
__inline Word32 sub(Word32 var1, Word32 var2)
|
|
{
|
|
#if ARMV5TE_SUB
|
|
Word32 result;
|
|
Word32 tmp;
|
|
asm (
|
|
"SUB %[result], %[var1], %[var2] \n"
|
|
"MOV %[tmp], %[var1], ASR #15 \n"
|
|
"TEQ %[tmp], %[var1], ASR #31 \n"
|
|
"EORNE %[result], %[mask], %[result], ASR #31 \n"
|
|
:[result]"=&r"(result), [tmp]"=&r"(tmp)
|
|
:[var1]"r"(var1), [var2]"r"(var2), [mask]"r"(0x7fff)
|
|
);
|
|
return result;
|
|
#else
|
|
Word32 var_out;
|
|
Word32 L_diff;
|
|
|
|
L_diff = (Word32) var1 - var2;
|
|
var_out = saturate(L_diff);
|
|
|
|
return (var_out);
|
|
#endif
|
|
}
|
|
#endif
|
|
|
|
/* Short division, 18 */
|
|
#if (DIV_S_IS_INLINE)
|
|
__inline Word32 div_s (Word32 var1, Word32 var2)
|
|
{
|
|
Word32 var_out = 0;
|
|
Word32 iteration;
|
|
Word32 L_num;
|
|
Word32 L_denom;
|
|
|
|
var_out = MAX_16;
|
|
if (var1!= var2)//var1!= var2
|
|
{
|
|
var_out = 0;
|
|
L_num = (Word32) var1;
|
|
|
|
L_denom = (Word32) var2;
|
|
|
|
//return (L_num<<15)/var2;
|
|
|
|
for (iteration = 0; iteration < 15; iteration++)
|
|
{
|
|
var_out <<= 1;
|
|
L_num <<= 1;
|
|
|
|
if (L_num >= L_denom)
|
|
{
|
|
L_num -= L_denom;
|
|
var_out++;
|
|
}
|
|
}
|
|
}
|
|
return (var_out);
|
|
}
|
|
#endif
|
|
|
|
/* Short mult, 1 */
|
|
#if (MULT_IS_INLINE)
|
|
__inline Word32 mult (Word32 var1, Word32 var2)
|
|
{
|
|
#if ARMV5TE_MULT && ARMV6_SAT
|
|
Word32 result;
|
|
asm (
|
|
"SMULBB %[result], %[var1], %[var2] \n"
|
|
"SSAT %[result], #16, %[result], ASR #15 \n"
|
|
:[result]"=r"(result)
|
|
:[var1]"r"(var1), [var2]"r"(var2)
|
|
);
|
|
return result;
|
|
#elif ARMV5TE_MULT
|
|
Word32 result, tmp;
|
|
asm (
|
|
"SMULBB %[tmp], %[var1], %[var2] \n"
|
|
"MOV %[result], %[tmp], ASR #15\n"
|
|
"MOV %[tmp], %[result], ASR #15\n"
|
|
"TEQ %[tmp], %[result], ASR #31\n"
|
|
"EORNE %[result], %[mask], %[result], ASR #31 \n"
|
|
:[result]"=&r"(result), [tmp]"=&r"(tmp)
|
|
:[var1]"r"(var1), [var2]"r"(var2), [mask]"r"(0x7fff)
|
|
);
|
|
return result;
|
|
#else
|
|
Word32 var_out;
|
|
Word32 L_product;
|
|
|
|
L_product = (Word32) var1 *(Word32) var2;
|
|
L_product = (L_product & (Word32) 0xffff8000L) >> 15;
|
|
if (L_product & (Word32) 0x00010000L)
|
|
L_product = L_product | (Word32) 0xffff0000L;
|
|
var_out = saturate(L_product);
|
|
|
|
return (var_out);
|
|
#endif
|
|
}
|
|
#endif
|
|
|
|
|
|
/* Short norm, 15 */
|
|
#if (NORM_S_IS_INLINE)
|
|
__inline Word32 norm_s (Word32 var1)
|
|
{
|
|
#if ARMV5TE_NORM_S
|
|
Word32 result;
|
|
Word32 tmp;
|
|
asm (
|
|
"RSBS %[tmp], %[var1], #0 \n"
|
|
"CLZLT %[result], %[var1]\n"
|
|
"CLZGT %[result], %[tmp]\n"
|
|
"SUBNE %[result], %[result], #17\n"
|
|
"MOVEQ %[result], #0\n"
|
|
"CMP %[var1], #-1\n"
|
|
"MOVEQ %[result], #15\n"
|
|
:[result]"=&r"(result), [tmp]"=&r"(tmp)
|
|
:[var1]"r"(var1)
|
|
);
|
|
return result;
|
|
#else
|
|
Word32 var_out;
|
|
|
|
if (var1 == 0)
|
|
{
|
|
var_out = 0;
|
|
}
|
|
else
|
|
{
|
|
if (var1 == -1)
|
|
{
|
|
var_out = 15;
|
|
}
|
|
else
|
|
{
|
|
if (var1 < 0)
|
|
{
|
|
var1 = (Word16)~var1;
|
|
}
|
|
for (var_out = 0; var1 < 0x4000; var_out++)
|
|
{
|
|
var1 <<= 1;
|
|
}
|
|
}
|
|
}
|
|
return (var_out);
|
|
#endif
|
|
}
|
|
#endif
|
|
|
|
/* Long norm, 30 */
|
|
#if (NORM_L_IS_INLINE)
|
|
__inline Word32 norm_l (Word32 L_var1)
|
|
{
|
|
#if ARMV5TE_NORM_L
|
|
Word32 result;
|
|
asm volatile(
|
|
"CMP %[L_var1], #0\n"
|
|
"CLZNE %[result], %[L_var1]\n"
|
|
"SUBNE %[result], %[result], #1\n"
|
|
"MOVEQ %[result], #0\n"
|
|
:[result]"=r"(result)
|
|
:[L_var1]"r"(L_var1)
|
|
);
|
|
return result;
|
|
#else
|
|
//Word16 var_out;
|
|
|
|
//if (L_var1 == 0)
|
|
//{
|
|
// var_out = 0;
|
|
//}
|
|
//else
|
|
//{
|
|
// if (L_var1 == (Word32) 0xffffffffL)
|
|
// {
|
|
// var_out = 31;
|
|
// }
|
|
// else
|
|
// {
|
|
// if (L_var1 < 0)
|
|
// {
|
|
// L_var1 = ~L_var1;
|
|
// }
|
|
// for (var_out = 0; L_var1 < (Word32) 0x40000000L; var_out++)
|
|
// {
|
|
// L_var1 <<= 1;
|
|
// }
|
|
// }
|
|
//}
|
|
//return (var_out);
|
|
Word16 a16;
|
|
Word16 r = 0 ;
|
|
|
|
|
|
if ( L_var1 < 0 ) {
|
|
L_var1 = ~L_var1;
|
|
}
|
|
|
|
if (0 == (L_var1 & 0x7fff8000)) {
|
|
a16 = extract_l(L_var1);
|
|
r += 16;
|
|
|
|
if (0 == (a16 & 0x7f80)) {
|
|
r += 8;
|
|
|
|
if (0 == (a16 & 0x0078)) {
|
|
r += 4;
|
|
|
|
if (0 == (a16 & 0x0006)) {
|
|
r += 2;
|
|
|
|
if (0 == (a16 & 0x0001)) {
|
|
r += 1;
|
|
}
|
|
}
|
|
else {
|
|
|
|
if (0 == (a16 & 0x0004)) {
|
|
r += 1;
|
|
}
|
|
}
|
|
}
|
|
else {
|
|
|
|
if (0 == (a16 & 0x0060)) {
|
|
r += 2;
|
|
|
|
if (0 == (a16 & 0x0010)) {
|
|
r += 1;
|
|
}
|
|
}
|
|
else {
|
|
|
|
if (0 == (a16 & 0x0040)) {
|
|
r += 1;
|
|
}
|
|
}
|
|
}
|
|
}
|
|
else {
|
|
|
|
if (0 == (a16 & 0x7800)) {
|
|
r += 4;
|
|
|
|
if (0 == (a16 & 0x0600)) {
|
|
r += 2;
|
|
|
|
if (0 == (a16 & 0x0100)) {
|
|
r += 1;
|
|
}
|
|
}
|
|
else {
|
|
|
|
if (0 == (a16 & 0x0400)) {
|
|
r += 1;
|
|
}
|
|
}
|
|
}
|
|
else {
|
|
|
|
if (0 == (a16 & 0x6000)) {
|
|
r += 2;
|
|
|
|
if (0 == (a16 & 0x1000)) {
|
|
r += 1;
|
|
}
|
|
}
|
|
else {
|
|
|
|
if (0 == (a16 & 0x4000)) {
|
|
r += 1;
|
|
}
|
|
}
|
|
}
|
|
}
|
|
}
|
|
else {
|
|
a16 = extract_h(L_var1);
|
|
|
|
if (0 == (a16 & 0x7f80)) {
|
|
r += 8;
|
|
|
|
if (0 == (a16 & 0x0078)) {
|
|
r += 4 ;
|
|
|
|
if (0 == (a16 & 0x0006)) {
|
|
r += 2;
|
|
|
|
if (0 == (a16 & 0x0001)) {
|
|
r += 1;
|
|
}
|
|
}
|
|
else {
|
|
|
|
if (0 == (a16 & 0x0004)) {
|
|
r += 1;
|
|
}
|
|
}
|
|
}
|
|
else {
|
|
|
|
if (0 == (a16 & 0x0060)) {
|
|
r += 2;
|
|
|
|
if (0 == (a16 & 0x0010)) {
|
|
r += 1;
|
|
}
|
|
}
|
|
else {
|
|
|
|
if (0 == (a16 & 0x0040)) {
|
|
r += 1;
|
|
}
|
|
}
|
|
}
|
|
}
|
|
else {
|
|
|
|
if (0 == (a16 & 0x7800)) {
|
|
r += 4;
|
|
|
|
if (0 == (a16 & 0x0600)) {
|
|
r += 2;
|
|
|
|
if (0 == (a16 & 0x0100)) {
|
|
r += 1;
|
|
}
|
|
}
|
|
else {
|
|
|
|
if (0 == (a16 & 0x0400)) {
|
|
r += 1;
|
|
}
|
|
}
|
|
}
|
|
else {
|
|
|
|
if (0 == (a16 & 0x6000)) {
|
|
r += 2;
|
|
|
|
if (0 == (a16 & 0x1000)) {
|
|
r += 1;
|
|
}
|
|
}
|
|
else {
|
|
|
|
if (0 == (a16 & 0x4000)) {
|
|
return 1;
|
|
}
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
return r ;
|
|
#endif
|
|
}
|
|
#endif
|
|
|
|
/* Round, 1 */
|
|
#if (ROUND_IS_INLINE)
|
|
__inline Word32 round16(Word32 L_var1)
|
|
{
|
|
#if ARMV5TE_ROUND
|
|
Word32 result;
|
|
asm (
|
|
"QADD %[result], %[L_var1], %[bias]\n"
|
|
"MOV %[result], %[result], ASR #16 \n"
|
|
:[result]"=r"(result)
|
|
:[L_var1]"r"(L_var1), [bias]"r"(0x8000)
|
|
);
|
|
return result;
|
|
#else
|
|
Word32 var_out;
|
|
Word32 L_rounded;
|
|
|
|
L_rounded = L_add (L_var1, (Word32) 0x00008000L);
|
|
var_out = extract_h (L_rounded);
|
|
return (var_out);
|
|
#endif
|
|
}
|
|
#endif
|
|
|
|
/* Mac, 1 */
|
|
#if (L_MAC_IS_INLINE)
|
|
__inline Word32 L_mac (Word32 L_var3, Word32 var1, Word32 var2)
|
|
{
|
|
#if ARMV5TE_L_MAC
|
|
Word32 result;
|
|
asm (
|
|
"SMULBB %[result], %[var1], %[var2]\n"
|
|
"QDADD %[result], %[L_var3], %[result]\n"
|
|
:[result]"=&r"(result)
|
|
: [L_var3]"r"(L_var3), [var1]"r"(var1), [var2]"r"(var2)
|
|
);
|
|
return result;
|
|
#else
|
|
Word32 L_var_out;
|
|
Word32 L_product;
|
|
|
|
L_product = L_mult(var1, var2);
|
|
L_var_out = L_add (L_var3, L_product);
|
|
return (L_var_out);
|
|
#endif
|
|
}
|
|
#endif
|
|
|
|
#if (L_ADD_IS_INLINE)
|
|
__inline Word32 L_add (Word32 L_var1, Word32 L_var2)
|
|
{
|
|
#if ARMV5TE_L_ADD
|
|
Word32 result;
|
|
asm (
|
|
"QADD %[result], %[L_var1], %[L_var2]\n"
|
|
:[result]"=r"(result)
|
|
:[L_var1]"r"(L_var1), [L_var2]"r"(L_var2)
|
|
);
|
|
return result;
|
|
#else
|
|
Word32 L_var_out;
|
|
|
|
L_var_out = L_var1 + L_var2;
|
|
if (((L_var1 ^ L_var2) & MIN_32) == 0)
|
|
{
|
|
if ((L_var_out ^ L_var1) & MIN_32)
|
|
{
|
|
L_var_out = (L_var1 < 0) ? MIN_32 : MAX_32;
|
|
}
|
|
}
|
|
return (L_var_out);
|
|
#endif
|
|
}
|
|
#endif
|
|
|
|
|
|
|
|
#if (MULT_R_IS_INLINE)
|
|
__inline Word32 mult_r (Word32 var1, Word32 var2)
|
|
{
|
|
Word32 var_out;
|
|
Word32 L_product_arr;
|
|
|
|
L_product_arr = (Word32)var1 *(Word32)var2; /* product */
|
|
L_product_arr += (Word32)0x00004000L; /* round */
|
|
L_product_arr >>= 15; /* shift */
|
|
|
|
var_out = saturate(L_product_arr);
|
|
|
|
return (var_out);
|
|
}
|
|
#endif
|
|
|
|
#if (SHR_R_IS_INLINE)
|
|
__inline Word32 shr_r (Word32 var1, Word32 var2)
|
|
{
|
|
Word32 var_out;
|
|
|
|
if (var2 > 15)
|
|
{
|
|
var_out = 0;
|
|
}
|
|
else
|
|
{
|
|
var_out = shr(var1, var2);
|
|
|
|
if (var2 > 0)
|
|
{
|
|
if ((var1 & ((Word16) 1 << (var2 - 1))) != 0)
|
|
{
|
|
var_out++;
|
|
}
|
|
}
|
|
}
|
|
|
|
return (var_out);
|
|
}
|
|
#endif
|
|
|
|
#if (MAC_R_IS_INLINE)
|
|
__inline Word32 mac_r (Word32 L_var3, Word32 var1, Word32 var2)
|
|
{
|
|
Word32 var_out;
|
|
|
|
L_var3 = L_mac (L_var3, var1, var2);
|
|
var_out = (Word16)((L_var3 + 0x8000L) >> 16);
|
|
|
|
return (var_out);
|
|
}
|
|
#endif
|
|
|
|
#if (MSU_R_IS_INLINE)
|
|
__inline Word32 msu_r (Word32 L_var3, Word32 var1, Word32 var2)
|
|
{
|
|
Word32 var_out;
|
|
|
|
L_var3 = L_msu (L_var3, var1, var2);
|
|
var_out = (Word16)((L_var3 + 0x8000L) >> 16);
|
|
|
|
return (var_out);
|
|
}
|
|
#endif
|
|
|
|
#if (L_SHR_R_IS_INLINE)
|
|
__inline Word32 L_shr_r (Word32 L_var1, Word32 var2)
|
|
{
|
|
Word32 L_var_out;
|
|
|
|
if (var2 > 31)
|
|
{
|
|
L_var_out = 0;
|
|
}
|
|
else
|
|
{
|
|
L_var_out = L_shr(L_var1, var2);
|
|
|
|
if (var2 > 0)
|
|
{
|
|
if ((L_var1 & ((Word32) 1 << (var2 - 1))) != 0)
|
|
{
|
|
L_var_out++;
|
|
}
|
|
}
|
|
}
|
|
|
|
return (L_var_out);
|
|
}
|
|
#endif
|
|
|
|
#if (EXTRACT_H_IS_INLINE)
|
|
__inline Word32 extract_h (Word32 L_var1)
|
|
{
|
|
Word32 var_out;
|
|
|
|
var_out = (L_var1 >> 16);
|
|
|
|
return (var_out);
|
|
}
|
|
#endif
|
|
|
|
#if (EXTRACT_L_IS_INLINE)
|
|
__inline Word32 extract_l(Word32 L_var1)
|
|
{
|
|
return (Word16) L_var1;
|
|
}
|
|
#endif
|
|
|
|
#endif
|