yemai/AT32F403A_407_Firmware_Library
1
0
Fork 0
mirror of https://github.com/ArteryTek/AT32F403A_407_Firmware_Library.git synced 2026-05-21 09:22:19 +00:00
Code Issues Actions Packages Projects Releases Wiki Activity

update version to v2.0.6

Browse Source
This commit is contained in:
Artery-MCU
2022-01-21 15:43:43 +08:00
parent 4fd69ebc78
commit e1d3f6e2c9
2457 changed files with 68919 additions and 60424 deletions
Download Patch File Download Diff File Expand all files Collapse all files

1034
libraries/cmsis/cm4/core_support/arm_common_tables.h
View File

File diff suppressed because it is too large Load Diff

152
libraries/cmsis/cm4/core_support/arm_const_structs.h
View File

@@ -1,76 +1,76 @@
/* ----------------------------------------------------------------------
* Project: CMSIS DSP Library
* Title: arm_const_structs.h
* Description: Constant structs that are initialized for user convenience.
* For example, some can be given as arguments to the arm_cfft_f32() function.
*
* $Date: 27. January 2017
* $Revision: V.1.5.1
*
* Target Processor: Cortex-M cores
* -------------------------------------------------------------------- */
/*
* Copyright (C) 2010-2017 ARM Limited or its affiliates. All rights reserved.
*
* SPDX-License-Identifier: Apache-2.0
*
* 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
*
* 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.
*/
#ifndef _ARM_CONST_STRUCTS_H
#define _ARM_CONST_STRUCTS_H
#include "arm_math.h"
#include "arm_common_tables.h"
extern const arm_cfft_instance_f64 arm_cfft_sR_f64_len16;
extern const arm_cfft_instance_f64 arm_cfft_sR_f64_len32;
extern const arm_cfft_instance_f64 arm_cfft_sR_f64_len64;
extern const arm_cfft_instance_f64 arm_cfft_sR_f64_len128;
extern const arm_cfft_instance_f64 arm_cfft_sR_f64_len256;
extern const arm_cfft_instance_f64 arm_cfft_sR_f64_len512;
extern const arm_cfft_instance_f64 arm_cfft_sR_f64_len1024;
extern const arm_cfft_instance_f64 arm_cfft_sR_f64_len2048;
extern const arm_cfft_instance_f64 arm_cfft_sR_f64_len4096;
extern const arm_cfft_instance_f32 arm_cfft_sR_f32_len16;
extern const arm_cfft_instance_f32 arm_cfft_sR_f32_len32;
extern const arm_cfft_instance_f32 arm_cfft_sR_f32_len64;
extern const arm_cfft_instance_f32 arm_cfft_sR_f32_len128;
extern const arm_cfft_instance_f32 arm_cfft_sR_f32_len256;
extern const arm_cfft_instance_f32 arm_cfft_sR_f32_len512;
extern const arm_cfft_instance_f32 arm_cfft_sR_f32_len1024;
extern const arm_cfft_instance_f32 arm_cfft_sR_f32_len2048;
extern const arm_cfft_instance_f32 arm_cfft_sR_f32_len4096;
extern const arm_cfft_instance_q31 arm_cfft_sR_q31_len16;
extern const arm_cfft_instance_q31 arm_cfft_sR_q31_len32;
extern const arm_cfft_instance_q31 arm_cfft_sR_q31_len64;
extern const arm_cfft_instance_q31 arm_cfft_sR_q31_len128;
extern const arm_cfft_instance_q31 arm_cfft_sR_q31_len256;
extern const arm_cfft_instance_q31 arm_cfft_sR_q31_len512;
extern const arm_cfft_instance_q31 arm_cfft_sR_q31_len1024;
extern const arm_cfft_instance_q31 arm_cfft_sR_q31_len2048;
extern const arm_cfft_instance_q31 arm_cfft_sR_q31_len4096;
extern const arm_cfft_instance_q15 arm_cfft_sR_q15_len16;
extern const arm_cfft_instance_q15 arm_cfft_sR_q15_len32;
extern const arm_cfft_instance_q15 arm_cfft_sR_q15_len64;
extern const arm_cfft_instance_q15 arm_cfft_sR_q15_len128;
extern const arm_cfft_instance_q15 arm_cfft_sR_q15_len256;
extern const arm_cfft_instance_q15 arm_cfft_sR_q15_len512;
extern const arm_cfft_instance_q15 arm_cfft_sR_q15_len1024;
extern const arm_cfft_instance_q15 arm_cfft_sR_q15_len2048;
extern const arm_cfft_instance_q15 arm_cfft_sR_q15_len4096;
#endif
/* ----------------------------------------------------------------------
* Project: CMSIS DSP Library
* Title: arm_const_structs.h
* Description: Constant structs that are initialized for user convenience.
* For example, some can be given as arguments to the arm_cfft_f32() function.
*
* $Date: 27. January 2017
* $Revision: V.1.5.1
*
* Target Processor: Cortex-M cores
* -------------------------------------------------------------------- */
/*
* Copyright (C) 2010-2017 ARM Limited or its affiliates. All rights reserved.
*
* SPDX-License-Identifier: Apache-2.0
*
* 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
*
* 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.
*/
#ifndef _ARM_CONST_STRUCTS_H
#define _ARM_CONST_STRUCTS_H
#include "arm_math.h"
#include "arm_common_tables.h"
extern const arm_cfft_instance_f64 arm_cfft_sR_f64_len16;
extern const arm_cfft_instance_f64 arm_cfft_sR_f64_len32;
extern const arm_cfft_instance_f64 arm_cfft_sR_f64_len64;
extern const arm_cfft_instance_f64 arm_cfft_sR_f64_len128;
extern const arm_cfft_instance_f64 arm_cfft_sR_f64_len256;
extern const arm_cfft_instance_f64 arm_cfft_sR_f64_len512;
extern const arm_cfft_instance_f64 arm_cfft_sR_f64_len1024;
extern const arm_cfft_instance_f64 arm_cfft_sR_f64_len2048;
extern const arm_cfft_instance_f64 arm_cfft_sR_f64_len4096;
extern const arm_cfft_instance_f32 arm_cfft_sR_f32_len16;
extern const arm_cfft_instance_f32 arm_cfft_sR_f32_len32;
extern const arm_cfft_instance_f32 arm_cfft_sR_f32_len64;
extern const arm_cfft_instance_f32 arm_cfft_sR_f32_len128;
extern const arm_cfft_instance_f32 arm_cfft_sR_f32_len256;
extern const arm_cfft_instance_f32 arm_cfft_sR_f32_len512;
extern const arm_cfft_instance_f32 arm_cfft_sR_f32_len1024;
extern const arm_cfft_instance_f32 arm_cfft_sR_f32_len2048;
extern const arm_cfft_instance_f32 arm_cfft_sR_f32_len4096;
extern const arm_cfft_instance_q31 arm_cfft_sR_q31_len16;
extern const arm_cfft_instance_q31 arm_cfft_sR_q31_len32;
extern const arm_cfft_instance_q31 arm_cfft_sR_q31_len64;
extern const arm_cfft_instance_q31 arm_cfft_sR_q31_len128;
extern const arm_cfft_instance_q31 arm_cfft_sR_q31_len256;
extern const arm_cfft_instance_q31 arm_cfft_sR_q31_len512;
extern const arm_cfft_instance_q31 arm_cfft_sR_q31_len1024;
extern const arm_cfft_instance_q31 arm_cfft_sR_q31_len2048;
extern const arm_cfft_instance_q31 arm_cfft_sR_q31_len4096;
extern const arm_cfft_instance_q15 arm_cfft_sR_q15_len16;
extern const arm_cfft_instance_q15 arm_cfft_sR_q15_len32;
extern const arm_cfft_instance_q15 arm_cfft_sR_q15_len64;
extern const arm_cfft_instance_q15 arm_cfft_sR_q15_len128;
extern const arm_cfft_instance_q15 arm_cfft_sR_q15_len256;
extern const arm_cfft_instance_q15 arm_cfft_sR_q15_len512;
extern const arm_cfft_instance_q15 arm_cfft_sR_q15_len1024;
extern const arm_cfft_instance_q15 arm_cfft_sR_q15_len2048;
extern const arm_cfft_instance_q15 arm_cfft_sR_q15_len4096;
#endif

696
libraries/cmsis/cm4/core_support/arm_helium_utils.h
View File

@@ -1,348 +1,348 @@
/* ----------------------------------------------------------------------
* Project: CMSIS DSP Library
* Title: arm_helium_utils.h
* Description: Utility functions for Helium development
*
* $Date: 09. September 2019
* $Revision: V.1.5.1
*
* Target Processor: Cortex-M cores
* -------------------------------------------------------------------- */
/*
* Copyright (C) 2010-2019 ARM Limited or its affiliates. All rights reserved.
*
* SPDX-License-Identifier: Apache-2.0
*
* 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
*
* 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.
*/
#ifndef _ARM_UTILS_HELIUM_H_
#define _ARM_UTILS_HELIUM_H_
/***************************************
Definitions available for MVEF and MVEI
***************************************/
#if defined (ARM_MATH_HELIUM) || defined(ARM_MATH_MVEF) || defined(ARM_MATH_MVEI)
#define INACTIVELANE 0 /* inactive lane content */
#endif /* defined (ARM_MATH_HELIUM) || defined(ARM_MATH_MVEF) || defined(ARM_MATH_MVEI) */
/***************************************
Definitions available for MVEF only
***************************************/
#if defined (ARM_MATH_HELIUM) || defined(ARM_MATH_MVEF)
__STATIC_FORCEINLINE float32_t vecAddAcrossF32Mve(float32x4_t in)
{
float32_t acc;
acc = vgetq_lane(in, 0) + vgetq_lane(in, 1) +
vgetq_lane(in, 2) + vgetq_lane(in, 3);
return acc;
}
/* newton initial guess */
#define INVSQRT_MAGIC_F32 0x5f3759df
#define INVSQRT_NEWTON_MVE_F32(invSqrt, xHalf, xStart)\
{ \
float32x4_t tmp; \
\
/* tmp = xhalf * x * x */ \
tmp = vmulq(xStart, xStart); \
tmp = vmulq(tmp, xHalf); \
/* (1.5f - xhalf * x * x) */ \
tmp = vsubq(vdupq_n_f32(1.5f), tmp); \
/* x = x*(1.5f-xhalf*x*x); */ \
invSqrt = vmulq(tmp, xStart); \
}
#endif /* defined (ARM_MATH_HELIUM) || defined(ARM_MATH_MVEF) */
/***************************************
Definitions available for MVEI only
***************************************/
#if defined (ARM_MATH_HELIUM) || defined(ARM_MATH_MVEI)
#include "arm_common_tables.h"
/* Following functions are used to transpose matrix in f32 and q31 cases */
__STATIC_INLINE arm_status arm_mat_trans_32bit_2x2_mve(
uint32_t * pDataSrc,
uint32_t * pDataDest)
{
static const uint32x4_t vecOffs = { 0, 2, 1, 3 };
/*
*
* | 0 1 | => | 0 2 |
* | 2 3 | | 1 3 |
*
*/
uint32x4_t vecIn = vldrwq_u32((uint32_t const *)pDataSrc);
vstrwq_scatter_shifted_offset_u32(pDataDest, vecOffs, vecIn);
return (ARM_MATH_SUCCESS);
}
__STATIC_INLINE arm_status arm_mat_trans_32bit_3x3_mve(
uint32_t * pDataSrc,
uint32_t * pDataDest)
{
const uint32x4_t vecOffs1 = { 0, 3, 6, 1};
const uint32x4_t vecOffs2 = { 4, 7, 2, 5};
/*
*
* | 0 1 2 | | 0 3 6 | 4 x 32 flattened version | 0 3 6 1 |
* | 3 4 5 | => | 1 4 7 | => | 4 7 2 5 |
* | 6 7 8 | | 2 5 8 | (row major) | 8 . . . |
*
*/
uint32x4_t vecIn1 = vldrwq_u32((uint32_t const *) pDataSrc);
uint32x4_t vecIn2 = vldrwq_u32((uint32_t const *) &pDataSrc[4]);
vstrwq_scatter_shifted_offset_u32(pDataDest, vecOffs1, vecIn1);
vstrwq_scatter_shifted_offset_u32(pDataDest, vecOffs2, vecIn2);
pDataDest[8] = pDataSrc[8];
return (ARM_MATH_SUCCESS);
}
__STATIC_INLINE arm_status arm_mat_trans_32bit_4x4_mve(uint32_t * pDataSrc, uint32_t * pDataDest)
{
/*
* 4x4 Matrix transposition
* is 4 x de-interleave operation
*
* 0 1 2 3 0 4 8 12
* 4 5 6 7 1 5 9 13
* 8 9 10 11 2 6 10 14
* 12 13 14 15 3 7 11 15
*/
uint32x4x4_t vecIn;
vecIn = vld4q((uint32_t const *) pDataSrc);
vstrwq(pDataDest, vecIn.val[0]);
pDataDest += 4;
vstrwq(pDataDest, vecIn.val[1]);
pDataDest += 4;
vstrwq(pDataDest, vecIn.val[2]);
pDataDest += 4;
vstrwq(pDataDest, vecIn.val[3]);
return (ARM_MATH_SUCCESS);
}
__STATIC_INLINE arm_status arm_mat_trans_32bit_generic_mve(
uint16_t srcRows,
uint16_t srcCols,
uint32_t * pDataSrc,
uint32_t * pDataDest)
{
uint32x4_t vecOffs;
uint32_t i;
uint32_t blkCnt;
uint32_t const *pDataC;
uint32_t *pDataDestR;
uint32x4_t vecIn;
vecOffs = vidupq_u32((uint32_t)0, 1);
vecOffs = vecOffs * srcCols;
i = srcCols;
do
{
pDataC = (uint32_t const *) pDataSrc;
pDataDestR = pDataDest;
blkCnt = srcRows >> 2;
while (blkCnt > 0U)
{
vecIn = vldrwq_gather_shifted_offset_u32(pDataC, vecOffs);
vstrwq(pDataDestR, vecIn);
pDataDestR += 4;
pDataC = pDataC + srcCols * 4;
/*
* Decrement the blockSize loop counter
*/
blkCnt--;
}
/*
* tail
*/
blkCnt = srcRows & 3;
if (blkCnt > 0U)
{
mve_pred16_t p0 = vctp32q(blkCnt);
vecIn = vldrwq_gather_shifted_offset_u32(pDataC, vecOffs);
vstrwq_p(pDataDestR, vecIn, p0);
}
pDataSrc += 1;
pDataDest += srcRows;
}
while (--i);
return (ARM_MATH_SUCCESS);
}
#if !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_ALL_FAST_TABLES) || defined(ARM_TABLE_FAST_SQRT_Q31_MVE)
__STATIC_INLINE q31x4_t FAST_VSQRT_Q31(q31x4_t vecIn)
{
q63x2_t vecTmpLL;
q31x4_t vecTmp0, vecTmp1;
q31_t scale;
q63_t tmp64;
q31x4_t vecNrm, vecDst, vecIdx, vecSignBits;
vecSignBits = vclsq(vecIn);
vecSignBits = vbicq(vecSignBits, 1);
/*
* in = in << no_of_sign_bits;
*/
vecNrm = vshlq(vecIn, vecSignBits);
/*
* index = in >> 24;
*/
vecIdx = vecNrm >> 24;
vecIdx = vecIdx << 1;
vecTmp0 = vldrwq_gather_shifted_offset_s32(sqrtTable_Q31, vecIdx);
vecIdx = vecIdx + 1;
vecTmp1 = vldrwq_gather_shifted_offset_s32(sqrtTable_Q31, vecIdx);
vecTmp1 = vqrdmulhq(vecTmp1, vecNrm);
vecTmp0 = vecTmp0 - vecTmp1;
vecTmp1 = vqrdmulhq(vecTmp0, vecTmp0);
vecTmp1 = vqrdmulhq(vecNrm, vecTmp1);
vecTmp1 = vdupq_n_s32(0x18000000) - vecTmp1;
vecTmp0 = vqrdmulhq(vecTmp0, vecTmp1);
vecTmpLL = vmullbq_int(vecNrm, vecTmp0);
/*
* scale elements 0, 2
*/
scale = 26 + (vecSignBits[0] >> 1);
tmp64 = asrl(vecTmpLL[0], scale);
vecDst[0] = (q31_t) tmp64;
scale = 26 + (vecSignBits[2] >> 1);
tmp64 = asrl(vecTmpLL[1], scale);
vecDst[2] = (q31_t) tmp64;
vecTmpLL = vmulltq_int(vecNrm, vecTmp0);
/*
* scale elements 1, 3
*/
scale = 26 + (vecSignBits[1] >> 1);
tmp64 = asrl(vecTmpLL[0], scale);
vecDst[1] = (q31_t) tmp64;
scale = 26 + (vecSignBits[3] >> 1);
tmp64 = asrl(vecTmpLL[1], scale);
vecDst[3] = (q31_t) tmp64;
/*
* set negative values to 0
*/
vecDst = vdupq_m(vecDst, 0, vcmpltq_n_s32(vecIn, 0));
return vecDst;
}
#endif
#if !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_ALL_FAST_TABLES) || defined(ARM_TABLE_FAST_SQRT_Q15_MVE)
__STATIC_INLINE q15x8_t FAST_VSQRT_Q15(q15x8_t vecIn)
{
q31x4_t vecTmpLev, vecTmpLodd, vecSignL;
q15x8_t vecTmp0, vecTmp1;
q15x8_t vecNrm, vecDst, vecIdx, vecSignBits;
vecDst = vuninitializedq_s16();
vecSignBits = vclsq(vecIn);
vecSignBits = vbicq(vecSignBits, 1);
/*
* in = in << no_of_sign_bits;
*/
vecNrm = vshlq(vecIn, vecSignBits);
vecIdx = vecNrm >> 8;
vecIdx = vecIdx << 1;
vecTmp0 = vldrhq_gather_shifted_offset_s16(sqrtTable_Q15, vecIdx);
vecIdx = vecIdx + 1;
vecTmp1 = vldrhq_gather_shifted_offset_s16(sqrtTable_Q15, vecIdx);
vecTmp1 = vqrdmulhq(vecTmp1, vecNrm);
vecTmp0 = vecTmp0 - vecTmp1;
vecTmp1 = vqrdmulhq(vecTmp0, vecTmp0);
vecTmp1 = vqrdmulhq(vecNrm, vecTmp1);
vecTmp1 = vdupq_n_s16(0x1800) - vecTmp1;
vecTmp0 = vqrdmulhq(vecTmp0, vecTmp1);
vecSignBits = vecSignBits >> 1;
vecTmpLev = vmullbq_int(vecNrm, vecTmp0);
vecTmpLodd = vmulltq_int(vecNrm, vecTmp0);
vecTmp0 = vecSignBits + 10;
/*
* negate sign to apply register based vshl
*/
vecTmp0 = -vecTmp0;
/*
* shift even elements
*/
vecSignL = vmovlbq(vecTmp0);
vecTmpLev = vshlq(vecTmpLev, vecSignL);
/*
* shift odd elements
*/
vecSignL = vmovltq(vecTmp0);
vecTmpLodd = vshlq(vecTmpLodd, vecSignL);
/*
* merge and narrow odd and even parts
*/
vecDst = vmovnbq_s32(vecDst, vecTmpLev);
vecDst = vmovntq_s32(vecDst, vecTmpLodd);
/*
* set negative values to 0
*/
vecDst = vdupq_m(vecDst, 0, vcmpltq_n_s16(vecIn, 0));
return vecDst;
}
#endif
#endif /* defined (ARM_MATH_HELIUM) || defined(ARM_MATH_MVEI) */
#endif
/* ----------------------------------------------------------------------
* Project: CMSIS DSP Library
* Title: arm_helium_utils.h
* Description: Utility functions for Helium development
*
* $Date: 09. September 2019
* $Revision: V.1.5.1
*
* Target Processor: Cortex-M cores
* -------------------------------------------------------------------- */
/*
* Copyright (C) 2010-2019 ARM Limited or its affiliates. All rights reserved.
*
* SPDX-License-Identifier: Apache-2.0
*
* 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
*
* 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.
*/
#ifndef _ARM_UTILS_HELIUM_H_
#define _ARM_UTILS_HELIUM_H_
/***************************************
Definitions available for MVEF and MVEI
***************************************/
#if defined (ARM_MATH_HELIUM) || defined(ARM_MATH_MVEF) || defined(ARM_MATH_MVEI)
#define INACTIVELANE 0 /* inactive lane content */
#endif /* defined (ARM_MATH_HELIUM) || defined(ARM_MATH_MVEF) || defined(ARM_MATH_MVEI) */
/***************************************
Definitions available for MVEF only
***************************************/
#if defined (ARM_MATH_HELIUM) || defined(ARM_MATH_MVEF)
__STATIC_FORCEINLINE float32_t vecAddAcrossF32Mve(float32x4_t in)
{
float32_t acc;
acc = vgetq_lane(in, 0) + vgetq_lane(in, 1) +
vgetq_lane(in, 2) + vgetq_lane(in, 3);
return acc;
}
/* newton initial guess */
#define INVSQRT_MAGIC_F32 0x5f3759df
#define INVSQRT_NEWTON_MVE_F32(invSqrt, xHalf, xStart)\
{ \
float32x4_t tmp; \
\
/* tmp = xhalf * x * x */ \
tmp = vmulq(xStart, xStart); \
tmp = vmulq(tmp, xHalf); \
/* (1.5f - xhalf * x * x) */ \
tmp = vsubq(vdupq_n_f32(1.5f), tmp); \
/* x = x*(1.5f-xhalf*x*x); */ \
invSqrt = vmulq(tmp, xStart); \
}
#endif /* defined (ARM_MATH_HELIUM) || defined(ARM_MATH_MVEF) */
/***************************************
Definitions available for MVEI only
***************************************/
#if defined (ARM_MATH_HELIUM) || defined(ARM_MATH_MVEI)
#include "arm_common_tables.h"
/* Following functions are used to transpose matrix in f32 and q31 cases */
__STATIC_INLINE arm_status arm_mat_trans_32bit_2x2_mve(
uint32_t * pDataSrc,
uint32_t * pDataDest)
{
static const uint32x4_t vecOffs = { 0, 2, 1, 3 };
/*
*
* | 0 1 | => | 0 2 |
* | 2 3 | | 1 3 |
*
*/
uint32x4_t vecIn = vldrwq_u32((uint32_t const *)pDataSrc);
vstrwq_scatter_shifted_offset_u32(pDataDest, vecOffs, vecIn);
return (ARM_MATH_SUCCESS);
}
__STATIC_INLINE arm_status arm_mat_trans_32bit_3x3_mve(
uint32_t * pDataSrc,
uint32_t * pDataDest)
{
const uint32x4_t vecOffs1 = { 0, 3, 6, 1};
const uint32x4_t vecOffs2 = { 4, 7, 2, 5};
/*
*
* | 0 1 2 | | 0 3 6 | 4 x 32 flattened version | 0 3 6 1 |
* | 3 4 5 | => | 1 4 7 | => | 4 7 2 5 |
* | 6 7 8 | | 2 5 8 | (row major) | 8 . . . |
*
*/
uint32x4_t vecIn1 = vldrwq_u32((uint32_t const *) pDataSrc);
uint32x4_t vecIn2 = vldrwq_u32((uint32_t const *) &pDataSrc[4]);
vstrwq_scatter_shifted_offset_u32(pDataDest, vecOffs1, vecIn1);
vstrwq_scatter_shifted_offset_u32(pDataDest, vecOffs2, vecIn2);
pDataDest[8] = pDataSrc[8];
return (ARM_MATH_SUCCESS);
}
__STATIC_INLINE arm_status arm_mat_trans_32bit_4x4_mve(uint32_t * pDataSrc, uint32_t * pDataDest)
{
/*
* 4x4 Matrix transposition
* is 4 x de-interleave operation
*
* 0 1 2 3 0 4 8 12
* 4 5 6 7 1 5 9 13
* 8 9 10 11 2 6 10 14
* 12 13 14 15 3 7 11 15
*/
uint32x4x4_t vecIn;
vecIn = vld4q((uint32_t const *) pDataSrc);
vstrwq(pDataDest, vecIn.val[0]);
pDataDest += 4;
vstrwq(pDataDest, vecIn.val[1]);
pDataDest += 4;
vstrwq(pDataDest, vecIn.val[2]);
pDataDest += 4;
vstrwq(pDataDest, vecIn.val[3]);
return (ARM_MATH_SUCCESS);
}
__STATIC_INLINE arm_status arm_mat_trans_32bit_generic_mve(
uint16_t srcRows,
uint16_t srcCols,
uint32_t * pDataSrc,
uint32_t * pDataDest)
{
uint32x4_t vecOffs;
uint32_t i;
uint32_t blkCnt;
uint32_t const *pDataC;
uint32_t *pDataDestR;
uint32x4_t vecIn;
vecOffs = vidupq_u32((uint32_t)0, 1);
vecOffs = vecOffs * srcCols;
i = srcCols;
do
{
pDataC = (uint32_t const *) pDataSrc;
pDataDestR = pDataDest;
blkCnt = srcRows >> 2;
while (blkCnt > 0U)
{
vecIn = vldrwq_gather_shifted_offset_u32(pDataC, vecOffs);
vstrwq(pDataDestR, vecIn);
pDataDestR += 4;
pDataC = pDataC + srcCols * 4;
/*
* Decrement the blockSize loop counter
*/
blkCnt--;
}
/*
* tail
*/
blkCnt = srcRows & 3;
if (blkCnt > 0U)
{
mve_pred16_t p0 = vctp32q(blkCnt);
vecIn = vldrwq_gather_shifted_offset_u32(pDataC, vecOffs);
vstrwq_p(pDataDestR, vecIn, p0);
}
pDataSrc += 1;
pDataDest += srcRows;
}
while (--i);
return (ARM_MATH_SUCCESS);
}
#if !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_ALL_FAST_TABLES) || defined(ARM_TABLE_FAST_SQRT_Q31_MVE)
__STATIC_INLINE q31x4_t FAST_VSQRT_Q31(q31x4_t vecIn)
{
q63x2_t vecTmpLL;
q31x4_t vecTmp0, vecTmp1;
q31_t scale;
q63_t tmp64;
q31x4_t vecNrm, vecDst, vecIdx, vecSignBits;
vecSignBits = vclsq(vecIn);
vecSignBits = vbicq(vecSignBits, 1);
/*
* in = in << no_of_sign_bits;
*/
vecNrm = vshlq(vecIn, vecSignBits);
/*
* index = in >> 24;
*/
vecIdx = vecNrm >> 24;
vecIdx = vecIdx << 1;
vecTmp0 = vldrwq_gather_shifted_offset_s32(sqrtTable_Q31, vecIdx);
vecIdx = vecIdx + 1;
vecTmp1 = vldrwq_gather_shifted_offset_s32(sqrtTable_Q31, vecIdx);
vecTmp1 = vqrdmulhq(vecTmp1, vecNrm);
vecTmp0 = vecTmp0 - vecTmp1;
vecTmp1 = vqrdmulhq(vecTmp0, vecTmp0);
vecTmp1 = vqrdmulhq(vecNrm, vecTmp1);
vecTmp1 = vdupq_n_s32(0x18000000) - vecTmp1;
vecTmp0 = vqrdmulhq(vecTmp0, vecTmp1);
vecTmpLL = vmullbq_int(vecNrm, vecTmp0);
/*
* scale elements 0, 2
*/
scale = 26 + (vecSignBits[0] >> 1);
tmp64 = asrl(vecTmpLL[0], scale);
vecDst[0] = (q31_t) tmp64;
scale = 26 + (vecSignBits[2] >> 1);
tmp64 = asrl(vecTmpLL[1], scale);
vecDst[2] = (q31_t) tmp64;
vecTmpLL = vmulltq_int(vecNrm, vecTmp0);
/*
* scale elements 1, 3
*/
scale = 26 + (vecSignBits[1] >> 1);
tmp64 = asrl(vecTmpLL[0], scale);
vecDst[1] = (q31_t) tmp64;
scale = 26 + (vecSignBits[3] >> 1);
tmp64 = asrl(vecTmpLL[1], scale);
vecDst[3] = (q31_t) tmp64;
/*
* set negative values to 0
*/
vecDst = vdupq_m(vecDst, 0, vcmpltq_n_s32(vecIn, 0));
return vecDst;
}
#endif
#if !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_ALL_FAST_TABLES) || defined(ARM_TABLE_FAST_SQRT_Q15_MVE)
__STATIC_INLINE q15x8_t FAST_VSQRT_Q15(q15x8_t vecIn)
{
q31x4_t vecTmpLev, vecTmpLodd, vecSignL;
q15x8_t vecTmp0, vecTmp1;
q15x8_t vecNrm, vecDst, vecIdx, vecSignBits;
vecDst = vuninitializedq_s16();
vecSignBits = vclsq(vecIn);
vecSignBits = vbicq(vecSignBits, 1);
/*
* in = in << no_of_sign_bits;
*/
vecNrm = vshlq(vecIn, vecSignBits);
vecIdx = vecNrm >> 8;
vecIdx = vecIdx << 1;
vecTmp0 = vldrhq_gather_shifted_offset_s16(sqrtTable_Q15, vecIdx);
vecIdx = vecIdx + 1;
vecTmp1 = vldrhq_gather_shifted_offset_s16(sqrtTable_Q15, vecIdx);
vecTmp1 = vqrdmulhq(vecTmp1, vecNrm);
vecTmp0 = vecTmp0 - vecTmp1;
vecTmp1 = vqrdmulhq(vecTmp0, vecTmp0);
vecTmp1 = vqrdmulhq(vecNrm, vecTmp1);
vecTmp1 = vdupq_n_s16(0x1800) - vecTmp1;
vecTmp0 = vqrdmulhq(vecTmp0, vecTmp1);
vecSignBits = vecSignBits >> 1;
vecTmpLev = vmullbq_int(vecNrm, vecTmp0);
vecTmpLodd = vmulltq_int(vecNrm, vecTmp0);
vecTmp0 = vecSignBits + 10;
/*
* negate sign to apply register based vshl
*/
vecTmp0 = -vecTmp0;
/*
* shift even elements
*/
vecSignL = vmovlbq(vecTmp0);
vecTmpLev = vshlq(vecTmpLev, vecSignL);
/*
* shift odd elements
*/
vecSignL = vmovltq(vecTmp0);
vecTmpLodd = vshlq(vecTmpLodd, vecSignL);
/*
* merge and narrow odd and even parts
*/
vecDst = vmovnbq_s32(vecDst, vecTmpLev);
vecDst = vmovntq_s32(vecDst, vecTmpLodd);
/*
* set negative values to 0
*/
vecDst = vdupq_m(vecDst, 0, vcmpltq_n_s16(vecIn, 0));
return vecDst;
}
#endif
#endif /* defined (ARM_MATH_HELIUM) || defined(ARM_MATH_MVEI) */
#endif

17940
libraries/cmsis/cm4/core_support/arm_math.h
View File

File diff suppressed because it is too large Load Diff

470
libraries/cmsis/cm4/core_support/arm_mve_tables.h
View File

@@ -1,235 +1,235 @@
/* ----------------------------------------------------------------------
* Project: CMSIS DSP Library
* Title: arm_mve_tables.h
* Description: common tables like fft twiddle factors, Bitreverse, reciprocal etc
* used for MVE implementation only
*
* $Date: 08. January 2020
* $Revision: V1.7.0
*
* Target Processor: Cortex-M cores
* -------------------------------------------------------------------- */
/*
* Copyright (C) 2010-2020 ARM Limited or its affiliates. All rights reserved.
*
* SPDX-License-Identifier: Apache-2.0
*
* 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
*
* 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.
*/
#ifndef _ARM_MVE_TABLES_H
#define _ARM_MVE_TABLES_H
#include "arm_math.h"
#if defined(ARM_MATH_MVEF) && !defined(ARM_MATH_AUTOVECTORIZE)
#if !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_FFT_ALLOW_TABLES)
#if !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_ALL_FFT_TABLES) || defined(ARM_TABLE_TWIDDLECOEF_F32_16) || defined(ARM_TABLE_TWIDDLECOEF_F32_32)
extern uint32_t rearranged_twiddle_tab_stride1_arr_16_f32[2];
extern uint32_t rearranged_twiddle_tab_stride2_arr_16_f32[2];
extern uint32_t rearranged_twiddle_tab_stride3_arr_16_f32[2];
extern float32_t rearranged_twiddle_stride1_16_f32[8];
extern float32_t rearranged_twiddle_stride2_16_f32[8];
extern float32_t rearranged_twiddle_stride3_16_f32[8];
#endif
#if !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_ALL_FFT_TABLES) || defined(ARM_TABLE_TWIDDLECOEF_F32_64) || defined(ARM_TABLE_TWIDDLECOEF_F32_128)
extern uint32_t rearranged_twiddle_tab_stride1_arr_64_f32[3];
extern uint32_t rearranged_twiddle_tab_stride2_arr_64_f32[3];
extern uint32_t rearranged_twiddle_tab_stride3_arr_64_f32[3];
extern float32_t rearranged_twiddle_stride1_64_f32[40];
extern float32_t rearranged_twiddle_stride2_64_f32[40];
extern float32_t rearranged_twiddle_stride3_64_f32[40];
#endif
#if !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_ALL_FFT_TABLES) || defined(ARM_TABLE_TWIDDLECOEF_F32_256) || defined(ARM_TABLE_TWIDDLECOEF_F32_512)
extern uint32_t rearranged_twiddle_tab_stride1_arr_256_f32[4];
extern uint32_t rearranged_twiddle_tab_stride2_arr_256_f32[4];
extern uint32_t rearranged_twiddle_tab_stride3_arr_256_f32[4];
extern float32_t rearranged_twiddle_stride1_256_f32[168];
extern float32_t rearranged_twiddle_stride2_256_f32[168];
extern float32_t rearranged_twiddle_stride3_256_f32[168];
#endif
#if !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_ALL_FFT_TABLES) || defined(ARM_TABLE_TWIDDLECOEF_F32_1024) || defined(ARM_TABLE_TWIDDLECOEF_F32_2048)
extern uint32_t rearranged_twiddle_tab_stride1_arr_1024_f32[5];
extern uint32_t rearranged_twiddle_tab_stride2_arr_1024_f32[5];
extern uint32_t rearranged_twiddle_tab_stride3_arr_1024_f32[5];
extern float32_t rearranged_twiddle_stride1_1024_f32[680];
extern float32_t rearranged_twiddle_stride2_1024_f32[680];
extern float32_t rearranged_twiddle_stride3_1024_f32[680];
#endif
#if !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_ALL_FFT_TABLES) || defined(ARM_TABLE_TWIDDLECOEF_F32_4096) || defined(ARM_TABLE_TWIDDLECOEF_F32_8192)
extern uint32_t rearranged_twiddle_tab_stride1_arr_4096_f32[6];
extern uint32_t rearranged_twiddle_tab_stride2_arr_4096_f32[6];
extern uint32_t rearranged_twiddle_tab_stride3_arr_4096_f32[6];
extern float32_t rearranged_twiddle_stride1_4096_f32[2728];
extern float32_t rearranged_twiddle_stride2_4096_f32[2728];
extern float32_t rearranged_twiddle_stride3_4096_f32[2728];
#endif
#endif /* !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_FFT_ALLOW_TABLES) */
#endif /* defined(ARM_MATH_MVEF) && !defined(ARM_MATH_AUTOVECTORIZE) */
#if defined(ARM_MATH_MVEI)
#if !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_FFT_ALLOW_TABLES)
#if !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_ALL_FFT_TABLES) || defined(ARM_TABLE_TWIDDLECOEF_Q31_16) || defined(ARM_TABLE_TWIDDLECOEF_Q31_32)
extern uint32_t rearranged_twiddle_tab_stride1_arr_16_q31[2];
extern uint32_t rearranged_twiddle_tab_stride2_arr_16_q31[2];
extern uint32_t rearranged_twiddle_tab_stride3_arr_16_q31[2];
extern q31_t rearranged_twiddle_stride1_16_q31[8];
extern q31_t rearranged_twiddle_stride2_16_q31[8];
extern q31_t rearranged_twiddle_stride3_16_q31[8];
#endif
#if !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_ALL_FFT_TABLES) || defined(ARM_TABLE_TWIDDLECOEF_Q31_64) || defined(ARM_TABLE_TWIDDLECOEF_Q31_128)
extern uint32_t rearranged_twiddle_tab_stride1_arr_64_q31[3];
extern uint32_t rearranged_twiddle_tab_stride2_arr_64_q31[3];
extern uint32_t rearranged_twiddle_tab_stride3_arr_64_q31[3];
extern q31_t rearranged_twiddle_stride1_64_q31[40];
extern q31_t rearranged_twiddle_stride2_64_q31[40];
extern q31_t rearranged_twiddle_stride3_64_q31[40];
#endif
#if !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_ALL_FFT_TABLES) || defined(ARM_TABLE_TWIDDLECOEF_Q31_256) || defined(ARM_TABLE_TWIDDLECOEF_Q31_512)
extern uint32_t rearranged_twiddle_tab_stride1_arr_256_q31[4];
extern uint32_t rearranged_twiddle_tab_stride2_arr_256_q31[4];
extern uint32_t rearranged_twiddle_tab_stride3_arr_256_q31[4];
extern q31_t rearranged_twiddle_stride1_256_q31[168];
extern q31_t rearranged_twiddle_stride2_256_q31[168];
extern q31_t rearranged_twiddle_stride3_256_q31[168];
#endif
#if !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_ALL_FFT_TABLES) || defined(ARM_TABLE_TWIDDLECOEF_Q31_1024) || defined(ARM_TABLE_TWIDDLECOEF_Q31_2048)
extern uint32_t rearranged_twiddle_tab_stride1_arr_1024_q31[5];
extern uint32_t rearranged_twiddle_tab_stride2_arr_1024_q31[5];
extern uint32_t rearranged_twiddle_tab_stride3_arr_1024_q31[5];
extern q31_t rearranged_twiddle_stride1_1024_q31[680];
extern q31_t rearranged_twiddle_stride2_1024_q31[680];
extern q31_t rearranged_twiddle_stride3_1024_q31[680];
#endif
#if !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_ALL_FFT_TABLES) || defined(ARM_TABLE_TWIDDLECOEF_Q31_4096) || defined(ARM_TABLE_TWIDDLECOEF_Q31_8192)
extern uint32_t rearranged_twiddle_tab_stride1_arr_4096_q31[6];
extern uint32_t rearranged_twiddle_tab_stride2_arr_4096_q31[6];
extern uint32_t rearranged_twiddle_tab_stride3_arr_4096_q31[6];
extern q31_t rearranged_twiddle_stride1_4096_q31[2728];
extern q31_t rearranged_twiddle_stride2_4096_q31[2728];
extern q31_t rearranged_twiddle_stride3_4096_q31[2728];
#endif
#endif /* !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_FFT_ALLOW_TABLES) */
#endif /* defined(ARM_MATH_MVEI) */
#if defined(ARM_MATH_MVEI)
#if !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_FFT_ALLOW_TABLES)
#if !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_ALL_FFT_TABLES) || defined(ARM_TABLE_TWIDDLECOEF_Q15_16) || defined(ARM_TABLE_TWIDDLECOEF_Q15_32)
extern uint32_t rearranged_twiddle_tab_stride1_arr_16_q15[2];
extern uint32_t rearranged_twiddle_tab_stride2_arr_16_q15[2];
extern uint32_t rearranged_twiddle_tab_stride3_arr_16_q15[2];
extern q15_t rearranged_twiddle_stride1_16_q15[8];
extern q15_t rearranged_twiddle_stride2_16_q15[8];
extern q15_t rearranged_twiddle_stride3_16_q15[8];
#endif
#if !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_ALL_FFT_TABLES) || defined(ARM_TABLE_TWIDDLECOEF_Q15_64) || defined(ARM_TABLE_TWIDDLECOEF_Q15_128)
extern uint32_t rearranged_twiddle_tab_stride1_arr_64_q15[3];
extern uint32_t rearranged_twiddle_tab_stride2_arr_64_q15[3];
extern uint32_t rearranged_twiddle_tab_stride3_arr_64_q15[3];
extern q15_t rearranged_twiddle_stride1_64_q15[40];
extern q15_t rearranged_twiddle_stride2_64_q15[40];
extern q15_t rearranged_twiddle_stride3_64_q15[40];
#endif
#if !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_ALL_FFT_TABLES) || defined(ARM_TABLE_TWIDDLECOEF_Q15_256) || defined(ARM_TABLE_TWIDDLECOEF_Q15_512)
extern uint32_t rearranged_twiddle_tab_stride1_arr_256_q15[4];
extern uint32_t rearranged_twiddle_tab_stride2_arr_256_q15[4];
extern uint32_t rearranged_twiddle_tab_stride3_arr_256_q15[4];
extern q15_t rearranged_twiddle_stride1_256_q15[168];
extern q15_t rearranged_twiddle_stride2_256_q15[168];
extern q15_t rearranged_twiddle_stride3_256_q15[168];
#endif
#if !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_ALL_FFT_TABLES) || defined(ARM_TABLE_TWIDDLECOEF_Q15_1024) || defined(ARM_TABLE_TWIDDLECOEF_Q15_2048)
extern uint32_t rearranged_twiddle_tab_stride1_arr_1024_q15[5];
extern uint32_t rearranged_twiddle_tab_stride2_arr_1024_q15[5];
extern uint32_t rearranged_twiddle_tab_stride3_arr_1024_q15[5];
extern q15_t rearranged_twiddle_stride1_1024_q15[680];
extern q15_t rearranged_twiddle_stride2_1024_q15[680];
extern q15_t rearranged_twiddle_stride3_1024_q15[680];
#endif
#if !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_ALL_FFT_TABLES) || defined(ARM_TABLE_TWIDDLECOEF_Q15_4096) || defined(ARM_TABLE_TWIDDLECOEF_Q15_8192)
extern uint32_t rearranged_twiddle_tab_stride1_arr_4096_q15[6];
extern uint32_t rearranged_twiddle_tab_stride2_arr_4096_q15[6];
extern uint32_t rearranged_twiddle_tab_stride3_arr_4096_q15[6];
extern q15_t rearranged_twiddle_stride1_4096_q15[2728];
extern q15_t rearranged_twiddle_stride2_4096_q15[2728];
extern q15_t rearranged_twiddle_stride3_4096_q15[2728];
#endif
#endif /* !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_FFT_ALLOW_TABLES) */
#endif /* defined(ARM_MATH_MVEI) */
#if defined(ARM_MATH_MVEI)
#if !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_FFT_ALLOW_TABLES)
#endif /* !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_FFT_ALLOW_TABLES) */
#endif /* defined(ARM_MATH_MVEI) */
#endif /*_ARM_MVE_TABLES_H*/
/* ----------------------------------------------------------------------
* Project: CMSIS DSP Library
* Title: arm_mve_tables.h
* Description: common tables like fft twiddle factors, Bitreverse, reciprocal etc
* used for MVE implementation only
*
* $Date: 08. January 2020
* $Revision: V1.7.0
*
* Target Processor: Cortex-M cores
* -------------------------------------------------------------------- */
/*
* Copyright (C) 2010-2020 ARM Limited or its affiliates. All rights reserved.
*
* SPDX-License-Identifier: Apache-2.0
*
* 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
*
* 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.
*/
#ifndef _ARM_MVE_TABLES_H
#define _ARM_MVE_TABLES_H
#include "arm_math.h"
#if defined(ARM_MATH_MVEF) && !defined(ARM_MATH_AUTOVECTORIZE)
#if !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_FFT_ALLOW_TABLES)
#if !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_ALL_FFT_TABLES) || defined(ARM_TABLE_TWIDDLECOEF_F32_16) || defined(ARM_TABLE_TWIDDLECOEF_F32_32)
extern uint32_t rearranged_twiddle_tab_stride1_arr_16_f32[2];
extern uint32_t rearranged_twiddle_tab_stride2_arr_16_f32[2];
extern uint32_t rearranged_twiddle_tab_stride3_arr_16_f32[2];
extern float32_t rearranged_twiddle_stride1_16_f32[8];
extern float32_t rearranged_twiddle_stride2_16_f32[8];
extern float32_t rearranged_twiddle_stride3_16_f32[8];
#endif
#if !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_ALL_FFT_TABLES) || defined(ARM_TABLE_TWIDDLECOEF_F32_64) || defined(ARM_TABLE_TWIDDLECOEF_F32_128)
extern uint32_t rearranged_twiddle_tab_stride1_arr_64_f32[3];
extern uint32_t rearranged_twiddle_tab_stride2_arr_64_f32[3];
extern uint32_t rearranged_twiddle_tab_stride3_arr_64_f32[3];
extern float32_t rearranged_twiddle_stride1_64_f32[40];
extern float32_t rearranged_twiddle_stride2_64_f32[40];
extern float32_t rearranged_twiddle_stride3_64_f32[40];
#endif
#if !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_ALL_FFT_TABLES) || defined(ARM_TABLE_TWIDDLECOEF_F32_256) || defined(ARM_TABLE_TWIDDLECOEF_F32_512)
extern uint32_t rearranged_twiddle_tab_stride1_arr_256_f32[4];
extern uint32_t rearranged_twiddle_tab_stride2_arr_256_f32[4];
extern uint32_t rearranged_twiddle_tab_stride3_arr_256_f32[4];
extern float32_t rearranged_twiddle_stride1_256_f32[168];
extern float32_t rearranged_twiddle_stride2_256_f32[168];
extern float32_t rearranged_twiddle_stride3_256_f32[168];
#endif
#if !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_ALL_FFT_TABLES) || defined(ARM_TABLE_TWIDDLECOEF_F32_1024) || defined(ARM_TABLE_TWIDDLECOEF_F32_2048)
extern uint32_t rearranged_twiddle_tab_stride1_arr_1024_f32[5];
extern uint32_t rearranged_twiddle_tab_stride2_arr_1024_f32[5];
extern uint32_t rearranged_twiddle_tab_stride3_arr_1024_f32[5];
extern float32_t rearranged_twiddle_stride1_1024_f32[680];
extern float32_t rearranged_twiddle_stride2_1024_f32[680];
extern float32_t rearranged_twiddle_stride3_1024_f32[680];
#endif
#if !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_ALL_FFT_TABLES) || defined(ARM_TABLE_TWIDDLECOEF_F32_4096) || defined(ARM_TABLE_TWIDDLECOEF_F32_8192)
extern uint32_t rearranged_twiddle_tab_stride1_arr_4096_f32[6];
extern uint32_t rearranged_twiddle_tab_stride2_arr_4096_f32[6];
extern uint32_t rearranged_twiddle_tab_stride3_arr_4096_f32[6];
extern float32_t rearranged_twiddle_stride1_4096_f32[2728];
extern float32_t rearranged_twiddle_stride2_4096_f32[2728];
extern float32_t rearranged_twiddle_stride3_4096_f32[2728];
#endif
#endif /* !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_FFT_ALLOW_TABLES) */
#endif /* defined(ARM_MATH_MVEF) && !defined(ARM_MATH_AUTOVECTORIZE) */
#if defined(ARM_MATH_MVEI)
#if !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_FFT_ALLOW_TABLES)
#if !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_ALL_FFT_TABLES) || defined(ARM_TABLE_TWIDDLECOEF_Q31_16) || defined(ARM_TABLE_TWIDDLECOEF_Q31_32)
extern uint32_t rearranged_twiddle_tab_stride1_arr_16_q31[2];
extern uint32_t rearranged_twiddle_tab_stride2_arr_16_q31[2];
extern uint32_t rearranged_twiddle_tab_stride3_arr_16_q31[2];
extern q31_t rearranged_twiddle_stride1_16_q31[8];
extern q31_t rearranged_twiddle_stride2_16_q31[8];
extern q31_t rearranged_twiddle_stride3_16_q31[8];
#endif
#if !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_ALL_FFT_TABLES) || defined(ARM_TABLE_TWIDDLECOEF_Q31_64) || defined(ARM_TABLE_TWIDDLECOEF_Q31_128)
extern uint32_t rearranged_twiddle_tab_stride1_arr_64_q31[3];
extern uint32_t rearranged_twiddle_tab_stride2_arr_64_q31[3];
extern uint32_t rearranged_twiddle_tab_stride3_arr_64_q31[3];
extern q31_t rearranged_twiddle_stride1_64_q31[40];
extern q31_t rearranged_twiddle_stride2_64_q31[40];
extern q31_t rearranged_twiddle_stride3_64_q31[40];
#endif
#if !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_ALL_FFT_TABLES) || defined(ARM_TABLE_TWIDDLECOEF_Q31_256) || defined(ARM_TABLE_TWIDDLECOEF_Q31_512)
extern uint32_t rearranged_twiddle_tab_stride1_arr_256_q31[4];
extern uint32_t rearranged_twiddle_tab_stride2_arr_256_q31[4];
extern uint32_t rearranged_twiddle_tab_stride3_arr_256_q31[4];
extern q31_t rearranged_twiddle_stride1_256_q31[168];
extern q31_t rearranged_twiddle_stride2_256_q31[168];
extern q31_t rearranged_twiddle_stride3_256_q31[168];
#endif
#if !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_ALL_FFT_TABLES) || defined(ARM_TABLE_TWIDDLECOEF_Q31_1024) || defined(ARM_TABLE_TWIDDLECOEF_Q31_2048)
extern uint32_t rearranged_twiddle_tab_stride1_arr_1024_q31[5];
extern uint32_t rearranged_twiddle_tab_stride2_arr_1024_q31[5];
extern uint32_t rearranged_twiddle_tab_stride3_arr_1024_q31[5];
extern q31_t rearranged_twiddle_stride1_1024_q31[680];
extern q31_t rearranged_twiddle_stride2_1024_q31[680];
extern q31_t rearranged_twiddle_stride3_1024_q31[680];
#endif
#if !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_ALL_FFT_TABLES) || defined(ARM_TABLE_TWIDDLECOEF_Q31_4096) || defined(ARM_TABLE_TWIDDLECOEF_Q31_8192)
extern uint32_t rearranged_twiddle_tab_stride1_arr_4096_q31[6];
extern uint32_t rearranged_twiddle_tab_stride2_arr_4096_q31[6];
extern uint32_t rearranged_twiddle_tab_stride3_arr_4096_q31[6];
extern q31_t rearranged_twiddle_stride1_4096_q31[2728];
extern q31_t rearranged_twiddle_stride2_4096_q31[2728];
extern q31_t rearranged_twiddle_stride3_4096_q31[2728];
#endif
#endif /* !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_FFT_ALLOW_TABLES) */
#endif /* defined(ARM_MATH_MVEI) */
#if defined(ARM_MATH_MVEI)
#if !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_FFT_ALLOW_TABLES)
#if !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_ALL_FFT_TABLES) || defined(ARM_TABLE_TWIDDLECOEF_Q15_16) || defined(ARM_TABLE_TWIDDLECOEF_Q15_32)
extern uint32_t rearranged_twiddle_tab_stride1_arr_16_q15[2];
extern uint32_t rearranged_twiddle_tab_stride2_arr_16_q15[2];
extern uint32_t rearranged_twiddle_tab_stride3_arr_16_q15[2];
extern q15_t rearranged_twiddle_stride1_16_q15[8];
extern q15_t rearranged_twiddle_stride2_16_q15[8];
extern q15_t rearranged_twiddle_stride3_16_q15[8];
#endif
#if !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_ALL_FFT_TABLES) || defined(ARM_TABLE_TWIDDLECOEF_Q15_64) || defined(ARM_TABLE_TWIDDLECOEF_Q15_128)
extern uint32_t rearranged_twiddle_tab_stride1_arr_64_q15[3];
extern uint32_t rearranged_twiddle_tab_stride2_arr_64_q15[3];
extern uint32_t rearranged_twiddle_tab_stride3_arr_64_q15[3];
extern q15_t rearranged_twiddle_stride1_64_q15[40];
extern q15_t rearranged_twiddle_stride2_64_q15[40];
extern q15_t rearranged_twiddle_stride3_64_q15[40];
#endif
#if !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_ALL_FFT_TABLES) || defined(ARM_TABLE_TWIDDLECOEF_Q15_256) || defined(ARM_TABLE_TWIDDLECOEF_Q15_512)
extern uint32_t rearranged_twiddle_tab_stride1_arr_256_q15[4];
extern uint32_t rearranged_twiddle_tab_stride2_arr_256_q15[4];
extern uint32_t rearranged_twiddle_tab_stride3_arr_256_q15[4];
extern q15_t rearranged_twiddle_stride1_256_q15[168];
extern q15_t rearranged_twiddle_stride2_256_q15[168];
extern q15_t rearranged_twiddle_stride3_256_q15[168];
#endif
#if !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_ALL_FFT_TABLES) || defined(ARM_TABLE_TWIDDLECOEF_Q15_1024) || defined(ARM_TABLE_TWIDDLECOEF_Q15_2048)
extern uint32_t rearranged_twiddle_tab_stride1_arr_1024_q15[5];
extern uint32_t rearranged_twiddle_tab_stride2_arr_1024_q15[5];
extern uint32_t rearranged_twiddle_tab_stride3_arr_1024_q15[5];
extern q15_t rearranged_twiddle_stride1_1024_q15[680];
extern q15_t rearranged_twiddle_stride2_1024_q15[680];
extern q15_t rearranged_twiddle_stride3_1024_q15[680];
#endif
#if !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_ALL_FFT_TABLES) || defined(ARM_TABLE_TWIDDLECOEF_Q15_4096) || defined(ARM_TABLE_TWIDDLECOEF_Q15_8192)
extern uint32_t rearranged_twiddle_tab_stride1_arr_4096_q15[6];
extern uint32_t rearranged_twiddle_tab_stride2_arr_4096_q15[6];
extern uint32_t rearranged_twiddle_tab_stride3_arr_4096_q15[6];
extern q15_t rearranged_twiddle_stride1_4096_q15[2728];
extern q15_t rearranged_twiddle_stride2_4096_q15[2728];
extern q15_t rearranged_twiddle_stride3_4096_q15[2728];
#endif
#endif /* !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_FFT_ALLOW_TABLES) */
#endif /* defined(ARM_MATH_MVEI) */
#if defined(ARM_MATH_MVEI)
#if !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_FFT_ALLOW_TABLES)
#endif /* !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_FFT_ALLOW_TABLES) */
#endif /* defined(ARM_MATH_MVEI) */
#endif /*_ARM_MVE_TABLES_H*/

744
libraries/cmsis/cm4/core_support/arm_vec_math.h
View File

@@ -1,372 +1,372 @@
/******************************************************************************
* @file arm_vec_math.h
* @brief Public header file for CMSIS DSP Library
* @version V1.7.0
* @date 15. October 2019
******************************************************************************/
/*
* Copyright (c) 2010-2019 Arm Limited or its affiliates. All rights reserved.
*
* SPDX-License-Identifier: Apache-2.0
*
* 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
*
* 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.
*/
#ifndef _ARM_VEC_MATH_H
#define _ARM_VEC_MATH_H
#include "arm_math.h"
#include "arm_common_tables.h"
#include "arm_helium_utils.h"
#ifdef __cplusplus
extern "C"
{
#endif
#if (defined(ARM_MATH_MVEF) || defined(ARM_MATH_HELIUM)) && !defined(ARM_MATH_AUTOVECTORIZE)
#define INV_NEWTON_INIT_F32 0x7EF127EA
static const float32_t __logf_rng_f32=0.693147180f;
/* fast inverse approximation (3x newton) */
__STATIC_INLINE f32x4_t vrecip_medprec_f32(
f32x4_t x)
{
q31x4_t m;
f32x4_t b;
any32x4_t xinv;
f32x4_t ax = vabsq(x);
xinv.f = ax;
m = 0x3F800000 - (xinv.i & 0x7F800000);
xinv.i = xinv.i + m;
xinv.f = 1.41176471f - 0.47058824f * xinv.f;
xinv.i = xinv.i + m;
b = 2.0f - xinv.f * ax;
xinv.f = xinv.f * b;
b = 2.0f - xinv.f * ax;
xinv.f = xinv.f * b;
b = 2.0f - xinv.f * ax;
xinv.f = xinv.f * b;
xinv.f = vdupq_m(xinv.f, INFINITY, vcmpeqq(x, 0.0f));
/*
* restore sign
*/
xinv.f = vnegq_m(xinv.f, xinv.f, vcmpltq(x, 0.0f));
return xinv.f;
}
/* fast inverse approximation (4x newton) */
__STATIC_INLINE f32x4_t vrecip_hiprec_f32(
f32x4_t x)
{
q31x4_t m;
f32x4_t b;
any32x4_t xinv;
f32x4_t ax = vabsq(x);
xinv.f = ax;
m = 0x3F800000 - (xinv.i & 0x7F800000);
xinv.i = xinv.i + m;
xinv.f = 1.41176471f - 0.47058824f * xinv.f;
xinv.i = xinv.i + m;
b = 2.0f - xinv.f * ax;
xinv.f = xinv.f * b;
b = 2.0f - xinv.f * ax;
xinv.f = xinv.f * b;
b = 2.0f - xinv.f * ax;
xinv.f = xinv.f * b;
b = 2.0f - xinv.f * ax;
xinv.f = xinv.f * b;
xinv.f = vdupq_m(xinv.f, INFINITY, vcmpeqq(x, 0.0f));
/*
* restore sign
*/
xinv.f = vnegq_m(xinv.f, xinv.f, vcmpltq(x, 0.0f));
return xinv.f;
}
__STATIC_INLINE f32x4_t vdiv_f32(
f32x4_t num, f32x4_t den)
{
return vmulq(num, vrecip_hiprec_f32(den));
}
/**
@brief Single-precision taylor dev.
@param[in] x f32 quad vector input
@param[in] coeffs f32 quad vector coeffs
@return destination f32 quad vector
*/
__STATIC_INLINE f32x4_t vtaylor_polyq_f32(
f32x4_t x,
const float32_t * coeffs)
{
f32x4_t A = vfmasq(vdupq_n_f32(coeffs[4]), x, coeffs[0]);
f32x4_t B = vfmasq(vdupq_n_f32(coeffs[6]), x, coeffs[2]);
f32x4_t C = vfmasq(vdupq_n_f32(coeffs[5]), x, coeffs[1]);
f32x4_t D = vfmasq(vdupq_n_f32(coeffs[7]), x, coeffs[3]);
f32x4_t x2 = vmulq(x, x);
f32x4_t x4 = vmulq(x2, x2);
f32x4_t res = vfmaq(vfmaq_f32(A, B, x2), vfmaq_f32(C, D, x2), x4);
return res;
}
__STATIC_INLINE f32x4_t vmant_exp_f32(
f32x4_t x,
int32x4_t * e)
{
any32x4_t r;
int32x4_t n;
r.f = x;
n = r.i >> 23;
n = n - 127;
r.i = r.i - (n << 23);
*e = n;
return r.f;
}
__STATIC_INLINE f32x4_t vlogq_f32(f32x4_t vecIn)
{
q31x4_t vecExpUnBiased;
f32x4_t vecTmpFlt0, vecTmpFlt1;
f32x4_t vecAcc0, vecAcc1, vecAcc2, vecAcc3;
f32x4_t vecExpUnBiasedFlt;
/*
* extract exponent
*/
vecTmpFlt1 = vmant_exp_f32(vecIn, &vecExpUnBiased);
vecTmpFlt0 = vecTmpFlt1 * vecTmpFlt1;
/*
* a = (__logf_lut_f32[4] * r.f) + (__logf_lut_f32[0]);
*/
vecAcc0 = vdupq_n_f32(__logf_lut_f32[0]);
vecAcc0 = vfmaq(vecAcc0, vecTmpFlt1, __logf_lut_f32[4]);
/*
* b = (__logf_lut_f32[6] * r.f) + (__logf_lut_f32[2]);
*/
vecAcc1 = vdupq_n_f32(__logf_lut_f32[2]);
vecAcc1 = vfmaq(vecAcc1, vecTmpFlt1, __logf_lut_f32[6]);
/*
* c = (__logf_lut_f32[5] * r.f) + (__logf_lut_f32[1]);
*/
vecAcc2 = vdupq_n_f32(__logf_lut_f32[1]);
vecAcc2 = vfmaq(vecAcc2, vecTmpFlt1, __logf_lut_f32[5]);
/*
* d = (__logf_lut_f32[7] * r.f) + (__logf_lut_f32[3]);
*/
vecAcc3 = vdupq_n_f32(__logf_lut_f32[3]);
vecAcc3 = vfmaq(vecAcc3, vecTmpFlt1, __logf_lut_f32[7]);
/*
* a = a + b * xx;
*/
vecAcc0 = vfmaq(vecAcc0, vecAcc1, vecTmpFlt0);
/*
* c = c + d * xx;
*/
vecAcc2 = vfmaq(vecAcc2, vecAcc3, vecTmpFlt0);
/*
* xx = xx * xx;
*/
vecTmpFlt0 = vecTmpFlt0 * vecTmpFlt0;
vecExpUnBiasedFlt = vcvtq_f32_s32(vecExpUnBiased);
/*
* r.f = a + c * xx;
*/
vecAcc0 = vfmaq(vecAcc0, vecAcc2, vecTmpFlt0);
/*
* add exponent
* r.f = r.f + ((float32_t) m) * __logf_rng_f32;
*/
vecAcc0 = vfmaq(vecAcc0, vecExpUnBiasedFlt, __logf_rng_f32);
// set log0 down to -inf
vecAcc0 = vdupq_m(vecAcc0, -INFINITY, vcmpeqq(vecIn, 0.0f));
return vecAcc0;
}
__STATIC_INLINE f32x4_t vexpq_f32(
f32x4_t x)
{
// Perform range reduction [-log(2),log(2)]
int32x4_t m = vcvtq_s32_f32(vmulq_n_f32(x, 1.4426950408f));
f32x4_t val = vfmsq_f32(x, vcvtq_f32_s32(m), vdupq_n_f32(0.6931471805f));
// Polynomial Approximation
f32x4_t poly = vtaylor_polyq_f32(val, exp_tab);
// Reconstruct
poly = (f32x4_t) (vqaddq_s32((q31x4_t) (poly), vqshlq_n_s32(m, 23)));
poly = vdupq_m(poly, 0.0f, vcmpltq_n_s32(m, -126));
return poly;
}
__STATIC_INLINE f32x4_t arm_vec_exponent_f32(f32x4_t x, int32_t nb)
{
f32x4_t r = x;
nb--;
while (nb > 0) {
r = vmulq(r, x);
nb--;
}
return (r);
}
__STATIC_INLINE f32x4_t vrecip_f32(f32x4_t vecIn)
{
f32x4_t vecSx, vecW, vecTmp;
any32x4_t v;
vecSx = vabsq(vecIn);
v.f = vecIn;
v.i = vsubq(vdupq_n_s32(INV_NEWTON_INIT_F32), v.i);
vecW = vmulq(vecSx, v.f);
// v.f = v.f * (8 + w * (-28 + w * (56 + w * (-70 + w *(56 + w * (-28 + w * (8 - w)))))));
vecTmp = vsubq(vdupq_n_f32(8.0f), vecW);
vecTmp = vfmasq(vecW, vecTmp, -28.0f);
vecTmp = vfmasq(vecW, vecTmp, 56.0f);
vecTmp = vfmasq(vecW, vecTmp, -70.0f);
vecTmp = vfmasq(vecW, vecTmp, 56.0f);
vecTmp = vfmasq(vecW, vecTmp, -28.0f);
vecTmp = vfmasq(vecW, vecTmp, 8.0f);
v.f = vmulq(v.f, vecTmp);
v.f = vdupq_m(v.f, INFINITY, vcmpeqq(vecIn, 0.0f));
/*
* restore sign
*/
v.f = vnegq_m(v.f, v.f, vcmpltq(vecIn, 0.0f));
return v.f;
}
__STATIC_INLINE f32x4_t vtanhq_f32(
f32x4_t val)
{
f32x4_t x =
vminnmq_f32(vmaxnmq_f32(val, vdupq_n_f32(-10.f)), vdupq_n_f32(10.0f));
f32x4_t exp2x = vexpq_f32(vmulq_n_f32(x, 2.f));
f32x4_t num = vsubq_n_f32(exp2x, 1.f);
f32x4_t den = vaddq_n_f32(exp2x, 1.f);
f32x4_t tanh = vmulq_f32(num, vrecip_f32(den));
return tanh;
}
__STATIC_INLINE f32x4_t vpowq_f32(
f32x4_t val,
f32x4_t n)
{
return vexpq_f32(vmulq_f32(n, vlogq_f32(val)));
}
#endif /* (defined(ARM_MATH_MVEF) || defined(ARM_MATH_HELIUM)) && !defined(ARM_MATH_AUTOVECTORIZE)*/
#if (defined(ARM_MATH_MVEI) || defined(ARM_MATH_HELIUM))
#endif /* (defined(ARM_MATH_MVEI) || defined(ARM_MATH_HELIUM)) */
#if (defined(ARM_MATH_NEON) || defined(ARM_MATH_NEON_EXPERIMENTAL)) && !defined(ARM_MATH_AUTOVECTORIZE)
#include "NEMath.h"
/**
* @brief Vectorized integer exponentiation
* @param[in] x value
* @param[in] nb integer exponent >= 1
* @return x^nb
*
*/
__STATIC_INLINE float32x4_t arm_vec_exponent_f32(float32x4_t x, int32_t nb)
{
float32x4_t r = x;
nb --;
while(nb > 0)
{
r = vmulq_f32(r , x);
nb--;
}
return(r);
}
__STATIC_INLINE float32x4_t __arm_vec_sqrt_f32_neon(float32x4_t x)
{
float32x4_t x1 = vmaxq_f32(x, vdupq_n_f32(FLT_MIN));
float32x4_t e = vrsqrteq_f32(x1);
e = vmulq_f32(vrsqrtsq_f32(vmulq_f32(x1, e), e), e);
e = vmulq_f32(vrsqrtsq_f32(vmulq_f32(x1, e), e), e);
return vmulq_f32(x, e);
}
__STATIC_INLINE int16x8_t __arm_vec_sqrt_q15_neon(int16x8_t vec)
{
float32x4_t tempF;
int32x4_t tempHI,tempLO;
tempLO = vmovl_s16(vget_low_s16(vec));
tempF = vcvtq_n_f32_s32(tempLO,15);
tempF = __arm_vec_sqrt_f32_neon(tempF);
tempLO = vcvtq_n_s32_f32(tempF,15);
tempHI = vmovl_s16(vget_high_s16(vec));
tempF = vcvtq_n_f32_s32(tempHI,15);
tempF = __arm_vec_sqrt_f32_neon(tempF);
tempHI = vcvtq_n_s32_f32(tempF,15);
return(vcombine_s16(vqmovn_s32(tempLO),vqmovn_s32(tempHI)));
}
__STATIC_INLINE int32x4_t __arm_vec_sqrt_q31_neon(int32x4_t vec)
{
float32x4_t temp;
temp = vcvtq_n_f32_s32(vec,31);
temp = __arm_vec_sqrt_f32_neon(temp);
return(vcvtq_n_s32_f32(temp,31));
}
#endif /* (defined(ARM_MATH_NEON) || defined(ARM_MATH_NEON_EXPERIMENTAL)) && !defined(ARM_MATH_AUTOVECTORIZE) */
#ifdef __cplusplus
}
#endif
#endif /* _ARM_VEC_MATH_H */
/**
*
* End of file.
*/
/******************************************************************************
* @file arm_vec_math.h
* @brief Public header file for CMSIS DSP Library
* @version V1.7.0
* @date 15. October 2019
******************************************************************************/
/*
* Copyright (c) 2010-2019 Arm Limited or its affiliates. All rights reserved.
*
* SPDX-License-Identifier: Apache-2.0
*
* 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
*
* 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.
*/
#ifndef _ARM_VEC_MATH_H
#define _ARM_VEC_MATH_H
#include "arm_math.h"
#include "arm_common_tables.h"
#include "arm_helium_utils.h"
#ifdef __cplusplus
extern "C"
{
#endif
#if (defined(ARM_MATH_MVEF) || defined(ARM_MATH_HELIUM)) && !defined(ARM_MATH_AUTOVECTORIZE)
#define INV_NEWTON_INIT_F32 0x7EF127EA
static const float32_t __logf_rng_f32=0.693147180f;
/* fast inverse approximation (3x newton) */
__STATIC_INLINE f32x4_t vrecip_medprec_f32(
f32x4_t x)
{
q31x4_t m;
f32x4_t b;
any32x4_t xinv;
f32x4_t ax = vabsq(x);
xinv.f = ax;
m = 0x3F800000 - (xinv.i & 0x7F800000);
xinv.i = xinv.i + m;
xinv.f = 1.41176471f - 0.47058824f * xinv.f;
xinv.i = xinv.i + m;
b = 2.0f - xinv.f * ax;
xinv.f = xinv.f * b;
b = 2.0f - xinv.f * ax;
xinv.f = xinv.f * b;
b = 2.0f - xinv.f * ax;
xinv.f = xinv.f * b;
xinv.f = vdupq_m(xinv.f, INFINITY, vcmpeqq(x, 0.0f));
/*
* restore sign
*/
xinv.f = vnegq_m(xinv.f, xinv.f, vcmpltq(x, 0.0f));
return xinv.f;
}
/* fast inverse approximation (4x newton) */
__STATIC_INLINE f32x4_t vrecip_hiprec_f32(
f32x4_t x)
{
q31x4_t m;
f32x4_t b;
any32x4_t xinv;
f32x4_t ax = vabsq(x);
xinv.f = ax;
m = 0x3F800000 - (xinv.i & 0x7F800000);
xinv.i = xinv.i + m;
xinv.f = 1.41176471f - 0.47058824f * xinv.f;
xinv.i = xinv.i + m;
b = 2.0f - xinv.f * ax;
xinv.f = xinv.f * b;
b = 2.0f - xinv.f * ax;
xinv.f = xinv.f * b;
b = 2.0f - xinv.f * ax;
xinv.f = xinv.f * b;
b = 2.0f - xinv.f * ax;
xinv.f = xinv.f * b;
xinv.f = vdupq_m(xinv.f, INFINITY, vcmpeqq(x, 0.0f));
/*
* restore sign
*/
xinv.f = vnegq_m(xinv.f, xinv.f, vcmpltq(x, 0.0f));
return xinv.f;
}
__STATIC_INLINE f32x4_t vdiv_f32(
f32x4_t num, f32x4_t den)
{
return vmulq(num, vrecip_hiprec_f32(den));
}
/**
@brief Single-precision taylor dev.
@param[in] x f32 quad vector input
@param[in] coeffs f32 quad vector coeffs
@return destination f32 quad vector
*/
__STATIC_INLINE f32x4_t vtaylor_polyq_f32(
f32x4_t x,
const float32_t * coeffs)
{
f32x4_t A = vfmasq(vdupq_n_f32(coeffs[4]), x, coeffs[0]);
f32x4_t B = vfmasq(vdupq_n_f32(coeffs[6]), x, coeffs[2]);
f32x4_t C = vfmasq(vdupq_n_f32(coeffs[5]), x, coeffs[1]);
f32x4_t D = vfmasq(vdupq_n_f32(coeffs[7]), x, coeffs[3]);
f32x4_t x2 = vmulq(x, x);
f32x4_t x4 = vmulq(x2, x2);
f32x4_t res = vfmaq(vfmaq_f32(A, B, x2), vfmaq_f32(C, D, x2), x4);
return res;
}
__STATIC_INLINE f32x4_t vmant_exp_f32(
f32x4_t x,
int32x4_t * e)
{
any32x4_t r;
int32x4_t n;
r.f = x;
n = r.i >> 23;
n = n - 127;
r.i = r.i - (n << 23);
*e = n;
return r.f;
}
__STATIC_INLINE f32x4_t vlogq_f32(f32x4_t vecIn)
{
q31x4_t vecExpUnBiased;
f32x4_t vecTmpFlt0, vecTmpFlt1;
f32x4_t vecAcc0, vecAcc1, vecAcc2, vecAcc3;
f32x4_t vecExpUnBiasedFlt;
/*
* extract exponent
*/
vecTmpFlt1 = vmant_exp_f32(vecIn, &vecExpUnBiased);
vecTmpFlt0 = vecTmpFlt1 * vecTmpFlt1;
/*
* a = (__logf_lut_f32[4] * r.f) + (__logf_lut_f32[0]);
*/
vecAcc0 = vdupq_n_f32(__logf_lut_f32[0]);
vecAcc0 = vfmaq(vecAcc0, vecTmpFlt1, __logf_lut_f32[4]);
/*
* b = (__logf_lut_f32[6] * r.f) + (__logf_lut_f32[2]);
*/
vecAcc1 = vdupq_n_f32(__logf_lut_f32[2]);
vecAcc1 = vfmaq(vecAcc1, vecTmpFlt1, __logf_lut_f32[6]);
/*
* c = (__logf_lut_f32[5] * r.f) + (__logf_lut_f32[1]);
*/
vecAcc2 = vdupq_n_f32(__logf_lut_f32[1]);
vecAcc2 = vfmaq(vecAcc2, vecTmpFlt1, __logf_lut_f32[5]);
/*
* d = (__logf_lut_f32[7] * r.f) + (__logf_lut_f32[3]);
*/
vecAcc3 = vdupq_n_f32(__logf_lut_f32[3]);
vecAcc3 = vfmaq(vecAcc3, vecTmpFlt1, __logf_lut_f32[7]);
/*
* a = a + b * xx;
*/
vecAcc0 = vfmaq(vecAcc0, vecAcc1, vecTmpFlt0);
/*
* c = c + d * xx;
*/
vecAcc2 = vfmaq(vecAcc2, vecAcc3, vecTmpFlt0);
/*
* xx = xx * xx;
*/
vecTmpFlt0 = vecTmpFlt0 * vecTmpFlt0;
vecExpUnBiasedFlt = vcvtq_f32_s32(vecExpUnBiased);
/*
* r.f = a + c * xx;
*/
vecAcc0 = vfmaq(vecAcc0, vecAcc2, vecTmpFlt0);
/*
* add exponent
* r.f = r.f + ((float32_t) m) * __logf_rng_f32;
*/
vecAcc0 = vfmaq(vecAcc0, vecExpUnBiasedFlt, __logf_rng_f32);
// set log0 down to -inf
vecAcc0 = vdupq_m(vecAcc0, -INFINITY, vcmpeqq(vecIn, 0.0f));
return vecAcc0;
}
__STATIC_INLINE f32x4_t vexpq_f32(
f32x4_t x)
{
// Perform range reduction [-log(2),log(2)]
int32x4_t m = vcvtq_s32_f32(vmulq_n_f32(x, 1.4426950408f));
f32x4_t val = vfmsq_f32(x, vcvtq_f32_s32(m), vdupq_n_f32(0.6931471805f));
// Polynomial Approximation
f32x4_t poly = vtaylor_polyq_f32(val, exp_tab);
// Reconstruct
poly = (f32x4_t) (vqaddq_s32((q31x4_t) (poly), vqshlq_n_s32(m, 23)));
poly = vdupq_m(poly, 0.0f, vcmpltq_n_s32(m, -126));
return poly;
}
__STATIC_INLINE f32x4_t arm_vec_exponent_f32(f32x4_t x, int32_t nb)
{
f32x4_t r = x;
nb--;
while (nb > 0) {
r = vmulq(r, x);
nb--;
}
return (r);
}
__STATIC_INLINE f32x4_t vrecip_f32(f32x4_t vecIn)
{
f32x4_t vecSx, vecW, vecTmp;
any32x4_t v;
vecSx = vabsq(vecIn);
v.f = vecIn;
v.i = vsubq(vdupq_n_s32(INV_NEWTON_INIT_F32), v.i);
vecW = vmulq(vecSx, v.f);
// v.f = v.f * (8 + w * (-28 + w * (56 + w * (-70 + w *(56 + w * (-28 + w * (8 - w)))))));
vecTmp = vsubq(vdupq_n_f32(8.0f), vecW);
vecTmp = vfmasq(vecW, vecTmp, -28.0f);
vecTmp = vfmasq(vecW, vecTmp, 56.0f);
vecTmp = vfmasq(vecW, vecTmp, -70.0f);
vecTmp = vfmasq(vecW, vecTmp, 56.0f);
vecTmp = vfmasq(vecW, vecTmp, -28.0f);
vecTmp = vfmasq(vecW, vecTmp, 8.0f);
v.f = vmulq(v.f, vecTmp);
v.f = vdupq_m(v.f, INFINITY, vcmpeqq(vecIn, 0.0f));
/*
* restore sign
*/
v.f = vnegq_m(v.f, v.f, vcmpltq(vecIn, 0.0f));
return v.f;
}
__STATIC_INLINE f32x4_t vtanhq_f32(
f32x4_t val)
{
f32x4_t x =
vminnmq_f32(vmaxnmq_f32(val, vdupq_n_f32(-10.f)), vdupq_n_f32(10.0f));
f32x4_t exp2x = vexpq_f32(vmulq_n_f32(x, 2.f));
f32x4_t num = vsubq_n_f32(exp2x, 1.f);
f32x4_t den = vaddq_n_f32(exp2x, 1.f);
f32x4_t tanh = vmulq_f32(num, vrecip_f32(den));
return tanh;
}
__STATIC_INLINE f32x4_t vpowq_f32(
f32x4_t val,
f32x4_t n)
{
return vexpq_f32(vmulq_f32(n, vlogq_f32(val)));
}
#endif /* (defined(ARM_MATH_MVEF) || defined(ARM_MATH_HELIUM)) && !defined(ARM_MATH_AUTOVECTORIZE)*/
#if (defined(ARM_MATH_MVEI) || defined(ARM_MATH_HELIUM))
#endif /* (defined(ARM_MATH_MVEI) || defined(ARM_MATH_HELIUM)) */
#if (defined(ARM_MATH_NEON) || defined(ARM_MATH_NEON_EXPERIMENTAL)) && !defined(ARM_MATH_AUTOVECTORIZE)
#include "NEMath.h"
/**
* @brief Vectorized integer exponentiation
* @param[in] x value
* @param[in] nb integer exponent >= 1
* @return x^nb
*
*/
__STATIC_INLINE float32x4_t arm_vec_exponent_f32(float32x4_t x, int32_t nb)
{
float32x4_t r = x;
nb --;
while(nb > 0)
{
r = vmulq_f32(r , x);
nb--;
}
return(r);
}
__STATIC_INLINE float32x4_t __arm_vec_sqrt_f32_neon(float32x4_t x)
{
float32x4_t x1 = vmaxq_f32(x, vdupq_n_f32(FLT_MIN));
float32x4_t e = vrsqrteq_f32(x1);
e = vmulq_f32(vrsqrtsq_f32(vmulq_f32(x1, e), e), e);
e = vmulq_f32(vrsqrtsq_f32(vmulq_f32(x1, e), e), e);
return vmulq_f32(x, e);
}
__STATIC_INLINE int16x8_t __arm_vec_sqrt_q15_neon(int16x8_t vec)
{
float32x4_t tempF;
int32x4_t tempHI,tempLO;
tempLO = vmovl_s16(vget_low_s16(vec));
tempF = vcvtq_n_f32_s32(tempLO,15);
tempF = __arm_vec_sqrt_f32_neon(tempF);
tempLO = vcvtq_n_s32_f32(tempF,15);
tempHI = vmovl_s16(vget_high_s16(vec));
tempF = vcvtq_n_f32_s32(tempHI,15);
tempF = __arm_vec_sqrt_f32_neon(tempF);
tempHI = vcvtq_n_s32_f32(tempF,15);
return(vcombine_s16(vqmovn_s32(tempLO),vqmovn_s32(tempHI)));
}
__STATIC_INLINE int32x4_t __arm_vec_sqrt_q31_neon(int32x4_t vec)
{
float32x4_t temp;
temp = vcvtq_n_f32_s32(vec,31);
temp = __arm_vec_sqrt_f32_neon(temp);
return(vcvtq_n_s32_f32(temp,31));
}
#endif /* (defined(ARM_MATH_NEON) || defined(ARM_MATH_NEON_EXPERIMENTAL)) && !defined(ARM_MATH_AUTOVECTORIZE) */
#ifdef __cplusplus
}
#endif
#endif /* _ARM_VEC_MATH_H */
/**
*
* End of file.
*/

1770
libraries/cmsis/cm4/core_support/cmsis_armcc.h
View File

File diff suppressed because it is too large Load Diff

2934
libraries/cmsis/cm4/core_support/cmsis_armclang.h
View File

File diff suppressed because it is too large Load Diff

3786
libraries/cmsis/cm4/core_support/cmsis_armclang_ltm.h
View File

File diff suppressed because it is too large Load Diff

4354
libraries/cmsis/cm4/core_support/cmsis_gcc.h
View File

File diff suppressed because it is too large Load Diff

704
libraries/cmsis/cm4/core_support/mpu_armv8.h
View File

@@ -1,352 +1,352 @@
/******************************************************************************
* @file mpu_armv8.h
* @brief CMSIS MPU API for Armv8-M and Armv8.1-M MPU
* @version V5.1.2
* @date 10. February 2020
******************************************************************************/
/*
* Copyright (c) 2017-2020 Arm Limited. All rights reserved.
*
* SPDX-License-Identifier: Apache-2.0
*
* 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
*
* 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.
*/
#if defined ( __ICCARM__ )
#pragma system_include /* treat file as system include file for MISRA check */
#elif defined (__clang__)
#pragma clang system_header /* treat file as system include file */
#endif
#ifndef ARM_MPU_ARMV8_H
#define ARM_MPU_ARMV8_H
/** \brief Attribute for device memory (outer only) */
#define ARM_MPU_ATTR_DEVICE ( 0U )
/** \brief Attribute for non-cacheable, normal memory */
#define ARM_MPU_ATTR_NON_CACHEABLE ( 4U )
/** \brief Attribute for normal memory (outer and inner)
* \param NT Non-Transient: Set to 1 for non-transient data.
* \param WB Write-Back: Set to 1 to use write-back update policy.
* \param RA Read Allocation: Set to 1 to use cache allocation on read miss.
* \param WA Write Allocation: Set to 1 to use cache allocation on write miss.
*/
#define ARM_MPU_ATTR_MEMORY_(NT, WB, RA, WA) \
((((NT) & 1U) << 3U) | (((WB) & 1U) << 2U) | (((RA) & 1U) << 1U) | ((WA) & 1U))
/** \brief Device memory type non Gathering, non Re-ordering, non Early Write Acknowledgement */
#define ARM_MPU_ATTR_DEVICE_nGnRnE (0U)
/** \brief Device memory type non Gathering, non Re-ordering, Early Write Acknowledgement */
#define ARM_MPU_ATTR_DEVICE_nGnRE (1U)
/** \brief Device memory type non Gathering, Re-ordering, Early Write Acknowledgement */
#define ARM_MPU_ATTR_DEVICE_nGRE (2U)
/** \brief Device memory type Gathering, Re-ordering, Early Write Acknowledgement */
#define ARM_MPU_ATTR_DEVICE_GRE (3U)
/** \brief Memory Attribute
* \param O Outer memory attributes
* \param I O == ARM_MPU_ATTR_DEVICE: Device memory attributes, else: Inner memory attributes
*/
#define ARM_MPU_ATTR(O, I) ((((O) & 0xFU) << 4U) | ((((O) & 0xFU) != 0U) ? ((I) & 0xFU) : (((I) & 0x3U) << 2U)))
/** \brief Normal memory non-shareable */
#define ARM_MPU_SH_NON (0U)
/** \brief Normal memory outer shareable */
#define ARM_MPU_SH_OUTER (2U)
/** \brief Normal memory inner shareable */
#define ARM_MPU_SH_INNER (3U)
/** \brief Memory access permissions
* \param RO Read-Only: Set to 1 for read-only memory.
* \param NP Non-Privileged: Set to 1 for non-privileged memory.
*/
#define ARM_MPU_AP_(RO, NP) ((((RO) & 1U) << 1U) | ((NP) & 1U))
/** \brief Region Base Address Register value
* \param BASE The base address bits [31:5] of a memory region. The value is zero extended. Effective address gets 32 byte aligned.
* \param SH Defines the Shareability domain for this memory region.
* \param RO Read-Only: Set to 1 for a read-only memory region.
* \param NP Non-Privileged: Set to 1 for a non-privileged memory region.
* \oaram XN eXecute Never: Set to 1 for a non-executable memory region.
*/
#define ARM_MPU_RBAR(BASE, SH, RO, NP, XN) \
(((BASE) & MPU_RBAR_BASE_Msk) | \
(((SH) << MPU_RBAR_SH_Pos) & MPU_RBAR_SH_Msk) | \
((ARM_MPU_AP_(RO, NP) << MPU_RBAR_AP_Pos) & MPU_RBAR_AP_Msk) | \
(((XN) << MPU_RBAR_XN_Pos) & MPU_RBAR_XN_Msk))
/** \brief Region Limit Address Register value
* \param LIMIT The limit address bits [31:5] for this memory region. The value is one extended.
* \param IDX The attribute index to be associated with this memory region.
*/
#define ARM_MPU_RLAR(LIMIT, IDX) \
(((LIMIT) & MPU_RLAR_LIMIT_Msk) | \
(((IDX) << MPU_RLAR_AttrIndx_Pos) & MPU_RLAR_AttrIndx_Msk) | \
(MPU_RLAR_EN_Msk))
#if defined(MPU_RLAR_PXN_Pos)
/** \brief Region Limit Address Register with PXN value
* \param LIMIT The limit address bits [31:5] for this memory region. The value is one extended.
* \param PXN Privileged execute never. Defines whether code can be executed from this privileged region.
* \param IDX The attribute index to be associated with this memory region.
*/
#define ARM_MPU_RLAR_PXN(LIMIT, PXN, IDX) \
(((LIMIT) & MPU_RLAR_LIMIT_Msk) | \
(((PXN) << MPU_RLAR_PXN_Pos) & MPU_RLAR_PXN_Msk) | \
(((IDX) << MPU_RLAR_AttrIndx_Pos) & MPU_RLAR_AttrIndx_Msk) | \
(MPU_RLAR_EN_Msk))
#endif
/**
* Struct for a single MPU Region
*/
typedef struct {
uint32_t RBAR; /*!< Region Base Address Register value */
uint32_t RLAR; /*!< Region Limit Address Register value */
} ARM_MPU_Region_t;
/** Enable the MPU.
* \param MPU_Control Default access permissions for unconfigured regions.
*/
__STATIC_INLINE void ARM_MPU_Enable(uint32_t MPU_Control)
{
__DMB();
MPU->CTRL = MPU_Control | MPU_CTRL_ENABLE_Msk;
#ifdef SCB_SHCSR_MEMFAULTENA_Msk
SCB->SHCSR |= SCB_SHCSR_MEMFAULTENA_Msk;
#endif
__DSB();
__ISB();
}
/** Disable the MPU.
*/
__STATIC_INLINE void ARM_MPU_Disable(void)
{
__DMB();
#ifdef SCB_SHCSR_MEMFAULTENA_Msk
SCB->SHCSR &= ~SCB_SHCSR_MEMFAULTENA_Msk;
#endif
MPU->CTRL &= ~MPU_CTRL_ENABLE_Msk;
__DSB();
__ISB();
}
#ifdef MPU_NS
/** Enable the Non-secure MPU.
* \param MPU_Control Default access permissions for unconfigured regions.
*/
__STATIC_INLINE void ARM_MPU_Enable_NS(uint32_t MPU_Control)
{
__DMB();
MPU_NS->CTRL = MPU_Control | MPU_CTRL_ENABLE_Msk;
#ifdef SCB_SHCSR_MEMFAULTENA_Msk
SCB_NS->SHCSR |= SCB_SHCSR_MEMFAULTENA_Msk;
#endif
__DSB();
__ISB();
}
/** Disable the Non-secure MPU.
*/
__STATIC_INLINE void ARM_MPU_Disable_NS(void)
{
__DMB();
#ifdef SCB_SHCSR_MEMFAULTENA_Msk
SCB_NS->SHCSR &= ~SCB_SHCSR_MEMFAULTENA_Msk;
#endif
MPU_NS->CTRL &= ~MPU_CTRL_ENABLE_Msk;
__DSB();
__ISB();
}
#endif
/** Set the memory attribute encoding to the given MPU.
* \param mpu Pointer to the MPU to be configured.
* \param idx The attribute index to be set [0-7]
* \param attr The attribute value to be set.
*/
__STATIC_INLINE void ARM_MPU_SetMemAttrEx(MPU_Type* mpu, uint8_t idx, uint8_t attr)
{
const uint8_t reg = idx / 4U;
const uint32_t pos = ((idx % 4U) * 8U);
const uint32_t mask = 0xFFU << pos;
if (reg >= (sizeof(mpu->MAIR) / sizeof(mpu->MAIR[0]))) {
return; // invalid index
}
mpu->MAIR[reg] = ((mpu->MAIR[reg] & ~mask) | ((attr << pos) & mask));
}
/** Set the memory attribute encoding.
* \param idx The attribute index to be set [0-7]
* \param attr The attribute value to be set.
*/
__STATIC_INLINE void ARM_MPU_SetMemAttr(uint8_t idx, uint8_t attr)
{
ARM_MPU_SetMemAttrEx(MPU, idx, attr);
}
#ifdef MPU_NS
/** Set the memory attribute encoding to the Non-secure MPU.
* \param idx The attribute index to be set [0-7]
* \param attr The attribute value to be set.
*/
__STATIC_INLINE void ARM_MPU_SetMemAttr_NS(uint8_t idx, uint8_t attr)
{
ARM_MPU_SetMemAttrEx(MPU_NS, idx, attr);
}
#endif
/** Clear and disable the given MPU region of the given MPU.
* \param mpu Pointer to MPU to be used.
* \param rnr Region number to be cleared.
*/
__STATIC_INLINE void ARM_MPU_ClrRegionEx(MPU_Type* mpu, uint32_t rnr)
{
mpu->RNR = rnr;
mpu->RLAR = 0U;
}
/** Clear and disable the given MPU region.
* \param rnr Region number to be cleared.
*/
__STATIC_INLINE void ARM_MPU_ClrRegion(uint32_t rnr)
{
ARM_MPU_ClrRegionEx(MPU, rnr);
}
#ifdef MPU_NS
/** Clear and disable the given Non-secure MPU region.
* \param rnr Region number to be cleared.
*/
__STATIC_INLINE void ARM_MPU_ClrRegion_NS(uint32_t rnr)
{
ARM_MPU_ClrRegionEx(MPU_NS, rnr);
}
#endif
/** Configure the given MPU region of the given MPU.
* \param mpu Pointer to MPU to be used.
* \param rnr Region number to be configured.
* \param rbar Value for RBAR register.
* \param rlar Value for RLAR register.
*/
__STATIC_INLINE void ARM_MPU_SetRegionEx(MPU_Type* mpu, uint32_t rnr, uint32_t rbar, uint32_t rlar)
{
mpu->RNR = rnr;
mpu->RBAR = rbar;
mpu->RLAR = rlar;
}
/** Configure the given MPU region.
* \param rnr Region number to be configured.
* \param rbar Value for RBAR register.
* \param rlar Value for RLAR register.
*/
__STATIC_INLINE void ARM_MPU_SetRegion(uint32_t rnr, uint32_t rbar, uint32_t rlar)
{
ARM_MPU_SetRegionEx(MPU, rnr, rbar, rlar);
}
#ifdef MPU_NS
/** Configure the given Non-secure MPU region.
* \param rnr Region number to be configured.
* \param rbar Value for RBAR register.
* \param rlar Value for RLAR register.
*/
__STATIC_INLINE void ARM_MPU_SetRegion_NS(uint32_t rnr, uint32_t rbar, uint32_t rlar)
{
ARM_MPU_SetRegionEx(MPU_NS, rnr, rbar, rlar);
}
#endif
/** Memcopy with strictly ordered memory access, e.g. for register targets.
* \param dst Destination data is copied to.
* \param src Source data is copied from.
* \param len Amount of data words to be copied.
*/
__STATIC_INLINE void ARM_MPU_OrderedMemcpy(volatile uint32_t* dst, const uint32_t* __RESTRICT src, uint32_t len)
{
uint32_t i;
for (i = 0U; i < len; ++i)
{
dst[i] = src[i];
}
}
/** Load the given number of MPU regions from a table to the given MPU.
* \param mpu Pointer to the MPU registers to be used.
* \param rnr First region number to be configured.
* \param table Pointer to the MPU configuration table.
* \param cnt Amount of regions to be configured.
*/
__STATIC_INLINE void ARM_MPU_LoadEx(MPU_Type* mpu, uint32_t rnr, ARM_MPU_Region_t const* table, uint32_t cnt)
{
const uint32_t rowWordSize = sizeof(ARM_MPU_Region_t)/4U;
if (cnt == 1U) {
mpu->RNR = rnr;
ARM_MPU_OrderedMemcpy(&(mpu->RBAR), &(table->RBAR), rowWordSize);
} else {
uint32_t rnrBase = rnr & ~(MPU_TYPE_RALIASES-1U);
uint32_t rnrOffset = rnr % MPU_TYPE_RALIASES;
mpu->RNR = rnrBase;
while ((rnrOffset + cnt) > MPU_TYPE_RALIASES) {
uint32_t c = MPU_TYPE_RALIASES - rnrOffset;
ARM_MPU_OrderedMemcpy(&(mpu->RBAR)+(rnrOffset*2U), &(table->RBAR), c*rowWordSize);
table += c;
cnt -= c;
rnrOffset = 0U;
rnrBase += MPU_TYPE_RALIASES;
mpu->RNR = rnrBase;
}
ARM_MPU_OrderedMemcpy(&(mpu->RBAR)+(rnrOffset*2U), &(table->RBAR), cnt*rowWordSize);
}
}
/** Load the given number of MPU regions from a table.
* \param rnr First region number to be configured.
* \param table Pointer to the MPU configuration table.
* \param cnt Amount of regions to be configured.
*/
__STATIC_INLINE void ARM_MPU_Load(uint32_t rnr, ARM_MPU_Region_t const* table, uint32_t cnt)
{
ARM_MPU_LoadEx(MPU, rnr, table, cnt);
}
#ifdef MPU_NS
/** Load the given number of MPU regions from a table to the Non-secure MPU.
* \param rnr First region number to be configured.
* \param table Pointer to the MPU configuration table.
* \param cnt Amount of regions to be configured.
*/
__STATIC_INLINE void ARM_MPU_Load_NS(uint32_t rnr, ARM_MPU_Region_t const* table, uint32_t cnt)
{
ARM_MPU_LoadEx(MPU_NS, rnr, table, cnt);
}
#endif
#endif
/******************************************************************************
* @file mpu_armv8.h
* @brief CMSIS MPU API for Armv8-M and Armv8.1-M MPU
* @version V5.1.2
* @date 10. February 2020
******************************************************************************/
/*
* Copyright (c) 2017-2020 Arm Limited. All rights reserved.
*
* SPDX-License-Identifier: Apache-2.0
*
* 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
*
* 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.
*/
#if defined ( __ICCARM__ )
#pragma system_include /* treat file as system include file for MISRA check */
#elif defined (__clang__)
#pragma clang system_header /* treat file as system include file */
#endif
#ifndef ARM_MPU_ARMV8_H
#define ARM_MPU_ARMV8_H
/** \brief Attribute for device memory (outer only) */
#define ARM_MPU_ATTR_DEVICE ( 0U )
/** \brief Attribute for non-cacheable, normal memory */
#define ARM_MPU_ATTR_NON_CACHEABLE ( 4U )
/** \brief Attribute for normal memory (outer and inner)
* \param NT Non-Transient: Set to 1 for non-transient data.
* \param WB Write-Back: Set to 1 to use write-back update policy.
* \param RA Read Allocation: Set to 1 to use cache allocation on read miss.
* \param WA Write Allocation: Set to 1 to use cache allocation on write miss.
*/
#define ARM_MPU_ATTR_MEMORY_(NT, WB, RA, WA) \
((((NT) & 1U) << 3U) | (((WB) & 1U) << 2U) | (((RA) & 1U) << 1U) | ((WA) & 1U))
/** \brief Device memory type non Gathering, non Re-ordering, non Early Write Acknowledgement */
#define ARM_MPU_ATTR_DEVICE_nGnRnE (0U)
/** \brief Device memory type non Gathering, non Re-ordering, Early Write Acknowledgement */
#define ARM_MPU_ATTR_DEVICE_nGnRE (1U)
/** \brief Device memory type non Gathering, Re-ordering, Early Write Acknowledgement */
#define ARM_MPU_ATTR_DEVICE_nGRE (2U)
/** \brief Device memory type Gathering, Re-ordering, Early Write Acknowledgement */
#define ARM_MPU_ATTR_DEVICE_GRE (3U)
/** \brief Memory Attribute
* \param O Outer memory attributes
* \param I O == ARM_MPU_ATTR_DEVICE: Device memory attributes, else: Inner memory attributes
*/
#define ARM_MPU_ATTR(O, I) ((((O) & 0xFU) << 4U) | ((((O) & 0xFU) != 0U) ? ((I) & 0xFU) : (((I) & 0x3U) << 2U)))
/** \brief Normal memory non-shareable */
#define ARM_MPU_SH_NON (0U)
/** \brief Normal memory outer shareable */
#define ARM_MPU_SH_OUTER (2U)
/** \brief Normal memory inner shareable */
#define ARM_MPU_SH_INNER (3U)
/** \brief Memory access permissions
* \param RO Read-Only: Set to 1 for read-only memory.
* \param NP Non-Privileged: Set to 1 for non-privileged memory.
*/
#define ARM_MPU_AP_(RO, NP) ((((RO) & 1U) << 1U) | ((NP) & 1U))
/** \brief Region Base Address Register value
* \param BASE The base address bits [31:5] of a memory region. The value is zero extended. Effective address gets 32 byte aligned.
* \param SH Defines the Shareability domain for this memory region.
* \param RO Read-Only: Set to 1 for a read-only memory region.
* \param NP Non-Privileged: Set to 1 for a non-privileged memory region.
* \oaram XN eXecute Never: Set to 1 for a non-executable memory region.
*/
#define ARM_MPU_RBAR(BASE, SH, RO, NP, XN) \
(((BASE) & MPU_RBAR_BASE_Msk) | \
(((SH) << MPU_RBAR_SH_Pos) & MPU_RBAR_SH_Msk) | \
((ARM_MPU_AP_(RO, NP) << MPU_RBAR_AP_Pos) & MPU_RBAR_AP_Msk) | \
(((XN) << MPU_RBAR_XN_Pos) & MPU_RBAR_XN_Msk))
/** \brief Region Limit Address Register value
* \param LIMIT The limit address bits [31:5] for this memory region. The value is one extended.
* \param IDX The attribute index to be associated with this memory region.
*/
#define ARM_MPU_RLAR(LIMIT, IDX) \
(((LIMIT) & MPU_RLAR_LIMIT_Msk) | \
(((IDX) << MPU_RLAR_AttrIndx_Pos) & MPU_RLAR_AttrIndx_Msk) | \
(MPU_RLAR_EN_Msk))
#if defined(MPU_RLAR_PXN_Pos)
/** \brief Region Limit Address Register with PXN value
* \param LIMIT The limit address bits [31:5] for this memory region. The value is one extended.
* \param PXN Privileged execute never. Defines whether code can be executed from this privileged region.
* \param IDX The attribute index to be associated with this memory region.
*/
#define ARM_MPU_RLAR_PXN(LIMIT, PXN, IDX) \
(((LIMIT) & MPU_RLAR_LIMIT_Msk) | \
(((PXN) << MPU_RLAR_PXN_Pos) & MPU_RLAR_PXN_Msk) | \
(((IDX) << MPU_RLAR_AttrIndx_Pos) & MPU_RLAR_AttrIndx_Msk) | \
(MPU_RLAR_EN_Msk))
#endif
/**
* Struct for a single MPU Region
*/
typedef struct {
uint32_t RBAR; /*!< Region Base Address Register value */
uint32_t RLAR; /*!< Region Limit Address Register value */
} ARM_MPU_Region_t;
/** Enable the MPU.
* \param MPU_Control Default access permissions for unconfigured regions.
*/
__STATIC_INLINE void ARM_MPU_Enable(uint32_t MPU_Control)
{
__DMB();
MPU->CTRL = MPU_Control | MPU_CTRL_ENABLE_Msk;
#ifdef SCB_SHCSR_MEMFAULTENA_Msk
SCB->SHCSR |= SCB_SHCSR_MEMFAULTENA_Msk;
#endif
__DSB();
__ISB();
}
/** Disable the MPU.
*/
__STATIC_INLINE void ARM_MPU_Disable(void)
{
__DMB();
#ifdef SCB_SHCSR_MEMFAULTENA_Msk
SCB->SHCSR &= ~SCB_SHCSR_MEMFAULTENA_Msk;
#endif
MPU->CTRL &= ~MPU_CTRL_ENABLE_Msk;
__DSB();
__ISB();
}
#ifdef MPU_NS
/** Enable the Non-secure MPU.
* \param MPU_Control Default access permissions for unconfigured regions.
*/
__STATIC_INLINE void ARM_MPU_Enable_NS(uint32_t MPU_Control)
{
__DMB();
MPU_NS->CTRL = MPU_Control | MPU_CTRL_ENABLE_Msk;
#ifdef SCB_SHCSR_MEMFAULTENA_Msk
SCB_NS->SHCSR |= SCB_SHCSR_MEMFAULTENA_Msk;
#endif
__DSB();
__ISB();
}
/** Disable the Non-secure MPU.
*/
__STATIC_INLINE void ARM_MPU_Disable_NS(void)
{
__DMB();
#ifdef SCB_SHCSR_MEMFAULTENA_Msk
SCB_NS->SHCSR &= ~SCB_SHCSR_MEMFAULTENA_Msk;
#endif
MPU_NS->CTRL &= ~MPU_CTRL_ENABLE_Msk;
__DSB();
__ISB();
}
#endif
/** Set the memory attribute encoding to the given MPU.
* \param mpu Pointer to the MPU to be configured.
* \param idx The attribute index to be set [0-7]
* \param attr The attribute value to be set.
*/
__STATIC_INLINE void ARM_MPU_SetMemAttrEx(MPU_Type* mpu, uint8_t idx, uint8_t attr)
{
const uint8_t reg = idx / 4U;
const uint32_t pos = ((idx % 4U) * 8U);
const uint32_t mask = 0xFFU << pos;
if (reg >= (sizeof(mpu->MAIR) / sizeof(mpu->MAIR[0]))) {
return; // invalid index
}
mpu->MAIR[reg] = ((mpu->MAIR[reg] & ~mask) | ((attr << pos) & mask));
}
/** Set the memory attribute encoding.
* \param idx The attribute index to be set [0-7]
* \param attr The attribute value to be set.
*/
__STATIC_INLINE void ARM_MPU_SetMemAttr(uint8_t idx, uint8_t attr)
{
ARM_MPU_SetMemAttrEx(MPU, idx, attr);
}
#ifdef MPU_NS
/** Set the memory attribute encoding to the Non-secure MPU.
* \param idx The attribute index to be set [0-7]
* \param attr The attribute value to be set.
*/
__STATIC_INLINE void ARM_MPU_SetMemAttr_NS(uint8_t idx, uint8_t attr)
{
ARM_MPU_SetMemAttrEx(MPU_NS, idx, attr);
}
#endif
/** Clear and disable the given MPU region of the given MPU.
* \param mpu Pointer to MPU to be used.
* \param rnr Region number to be cleared.
*/
__STATIC_INLINE void ARM_MPU_ClrRegionEx(MPU_Type* mpu, uint32_t rnr)
{
mpu->RNR = rnr;
mpu->RLAR = 0U;
}
/** Clear and disable the given MPU region.
* \param rnr Region number to be cleared.
*/
__STATIC_INLINE void ARM_MPU_ClrRegion(uint32_t rnr)
{
ARM_MPU_ClrRegionEx(MPU, rnr);
}
#ifdef MPU_NS
/** Clear and disable the given Non-secure MPU region.
* \param rnr Region number to be cleared.
*/
__STATIC_INLINE void ARM_MPU_ClrRegion_NS(uint32_t rnr)
{
ARM_MPU_ClrRegionEx(MPU_NS, rnr);
}
#endif
/** Configure the given MPU region of the given MPU.
* \param mpu Pointer to MPU to be used.
* \param rnr Region number to be configured.
* \param rbar Value for RBAR register.
* \param rlar Value for RLAR register.
*/
__STATIC_INLINE void ARM_MPU_SetRegionEx(MPU_Type* mpu, uint32_t rnr, uint32_t rbar, uint32_t rlar)
{
mpu->RNR = rnr;
mpu->RBAR = rbar;
mpu->RLAR = rlar;
}
/** Configure the given MPU region.
* \param rnr Region number to be configured.
* \param rbar Value for RBAR register.
* \param rlar Value for RLAR register.
*/
__STATIC_INLINE void ARM_MPU_SetRegion(uint32_t rnr, uint32_t rbar, uint32_t rlar)
{
ARM_MPU_SetRegionEx(MPU, rnr, rbar, rlar);
}
#ifdef MPU_NS
/** Configure the given Non-secure MPU region.
* \param rnr Region number to be configured.
* \param rbar Value for RBAR register.
* \param rlar Value for RLAR register.
*/
__STATIC_INLINE void ARM_MPU_SetRegion_NS(uint32_t rnr, uint32_t rbar, uint32_t rlar)
{
ARM_MPU_SetRegionEx(MPU_NS, rnr, rbar, rlar);
}
#endif
/** Memcopy with strictly ordered memory access, e.g. for register targets.
* \param dst Destination data is copied to.
* \param src Source data is copied from.
* \param len Amount of data words to be copied.
*/
__STATIC_INLINE void ARM_MPU_OrderedMemcpy(volatile uint32_t* dst, const uint32_t* __RESTRICT src, uint32_t len)
{
uint32_t i;
for (i = 0U; i < len; ++i)
{
dst[i] = src[i];
}
}
/** Load the given number of MPU regions from a table to the given MPU.
* \param mpu Pointer to the MPU registers to be used.
* \param rnr First region number to be configured.
* \param table Pointer to the MPU configuration table.
* \param cnt Amount of regions to be configured.
*/
__STATIC_INLINE void ARM_MPU_LoadEx(MPU_Type* mpu, uint32_t rnr, ARM_MPU_Region_t const* table, uint32_t cnt)
{
const uint32_t rowWordSize = sizeof(ARM_MPU_Region_t)/4U;
if (cnt == 1U) {
mpu->RNR = rnr;
ARM_MPU_OrderedMemcpy(&(mpu->RBAR), &(table->RBAR), rowWordSize);
} else {
uint32_t rnrBase = rnr & ~(MPU_TYPE_RALIASES-1U);
uint32_t rnrOffset = rnr % MPU_TYPE_RALIASES;
mpu->RNR = rnrBase;
while ((rnrOffset + cnt) > MPU_TYPE_RALIASES) {
uint32_t c = MPU_TYPE_RALIASES - rnrOffset;
ARM_MPU_OrderedMemcpy(&(mpu->RBAR)+(rnrOffset*2U), &(table->RBAR), c*rowWordSize);
table += c;
cnt -= c;
rnrOffset = 0U;
rnrBase += MPU_TYPE_RALIASES;
mpu->RNR = rnrBase;
}
ARM_MPU_OrderedMemcpy(&(mpu->RBAR)+(rnrOffset*2U), &(table->RBAR), cnt*rowWordSize);
}
}
/** Load the given number of MPU regions from a table.
* \param rnr First region number to be configured.
* \param table Pointer to the MPU configuration table.
* \param cnt Amount of regions to be configured.
*/
__STATIC_INLINE void ARM_MPU_Load(uint32_t rnr, ARM_MPU_Region_t const* table, uint32_t cnt)
{
ARM_MPU_LoadEx(MPU, rnr, table, cnt);
}
#ifdef MPU_NS
/** Load the given number of MPU regions from a table to the Non-secure MPU.
* \param rnr First region number to be configured.
* \param table Pointer to the MPU configuration table.
* \param cnt Amount of regions to be configured.
*/
__STATIC_INLINE void ARM_MPU_Load_NS(uint32_t rnr, ARM_MPU_Region_t const* table, uint32_t cnt)
{
ARM_MPU_LoadEx(MPU_NS, rnr, table, cnt);
}
#endif
#endif

674
libraries/cmsis/cm4/core_support/pmu_armv8.h
View File

@@ -1,337 +1,337 @@
/******************************************************************************
* @file pmu_armv8.h
* @brief CMSIS PMU API for Armv8.1-M PMU
* @version V1.0.0
* @date 24. March 2020
******************************************************************************/
/*
* Copyright (c) 2020 Arm Limited. All rights reserved.
*
* SPDX-License-Identifier: Apache-2.0
*
* 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
*
* 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.
*/
#if defined ( __ICCARM__ )
#pragma system_include /* treat file as system include file for MISRA check */
#elif defined (__clang__)
#pragma clang system_header /* treat file as system include file */
#endif
#ifndef ARM_PMU_ARMV8_H
#define ARM_PMU_ARMV8_H
/**
* \brief PMU Events
* \note See the Armv8.1-M Architecture Reference Manual for full details on these PMU events.
* */
#define ARM_PMU_SW_INCR 0x0000 /*!< Software update to the PMU_SWINC register, architecturally executed and condition code check pass */
#define ARM_PMU_L1I_CACHE_REFILL 0x0001 /*!< L1 I-Cache refill */
#define ARM_PMU_L1D_CACHE_REFILL 0x0003 /*!< L1 D-Cache refill */
#define ARM_PMU_L1D_CACHE 0x0004 /*!< L1 D-Cache access */
#define ARM_PMU_LD_RETIRED 0x0006 /*!< Memory-reading instruction architecturally executed and condition code check pass */
#define ARM_PMU_ST_RETIRED 0x0007 /*!< Memory-writing instruction architecturally executed and condition code check pass */
#define ARM_PMU_INST_RETIRED 0x0008 /*!< Instruction architecturally executed */
#define ARM_PMU_EXC_TAKEN 0x0009 /*!< Exception entry */
#define ARM_PMU_EXC_RETURN 0x000A /*!< Exception return instruction architecturally executed and the condition code check pass */
#define ARM_PMU_PC_WRITE_RETIRED 0x000C /*!< Software change to the Program Counter (PC). Instruction is architecturally executed and condition code check pass */
#define ARM_PMU_BR_IMMED_RETIRED 0x000D /*!< Immediate branch architecturally executed */
#define ARM_PMU_BR_RETURN_RETIRED 0x000E /*!< Function return instruction architecturally executed and the condition code check pass */
#define ARM_PMU_UNALIGNED_LDST_RETIRED 0x000F /*!< Unaligned memory memory-reading or memory-writing instruction architecturally executed and condition code check pass */
#define ARM_PMU_BR_MIS_PRED 0x0010 /*!< Mispredicted or not predicted branch speculatively executed */
#define ARM_PMU_CPU_CYCLES 0x0011 /*!< Cycle */
#define ARM_PMU_BR_PRED 0x0012 /*!< Predictable branch speculatively executed */
#define ARM_PMU_MEM_ACCESS 0x0013 /*!< Data memory access */
#define ARM_PMU_L1I_CACHE 0x0014 /*!< Level 1 instruction cache access */
#define ARM_PMU_L1D_CACHE_WB 0x0015 /*!< Level 1 data cache write-back */
#define ARM_PMU_L2D_CACHE 0x0016 /*!< Level 2 data cache access */
#define ARM_PMU_L2D_CACHE_REFILL 0x0017 /*!< Level 2 data cache refill */
#define ARM_PMU_L2D_CACHE_WB 0x0018 /*!< Level 2 data cache write-back */
#define ARM_PMU_BUS_ACCESS 0x0019 /*!< Bus access */
#define ARM_PMU_MEMORY_ERROR 0x001A /*!< Local memory error */
#define ARM_PMU_INST_SPEC 0x001B /*!< Instruction speculatively executed */
#define ARM_PMU_BUS_CYCLES 0x001D /*!< Bus cycles */
#define ARM_PMU_CHAIN 0x001E /*!< For an odd numbered counter, increment when an overflow occurs on the preceding even-numbered counter on the same PE */
#define ARM_PMU_L1D_CACHE_ALLOCATE 0x001F /*!< Level 1 data cache allocation without refill */
#define ARM_PMU_L2D_CACHE_ALLOCATE 0x0020 /*!< Level 2 data cache allocation without refill */
#define ARM_PMU_BR_RETIRED 0x0021 /*!< Branch instruction architecturally executed */
#define ARM_PMU_BR_MIS_PRED_RETIRED 0x0022 /*!< Mispredicted branch instruction architecturally executed */
#define ARM_PMU_STALL_FRONTEND 0x0023 /*!< No operation issued because of the frontend */
#define ARM_PMU_STALL_BACKEND 0x0024 /*!< No operation issued because of the backend */
#define ARM_PMU_L2I_CACHE 0x0027 /*!< Level 2 instruction cache access */
#define ARM_PMU_L2I_CACHE_REFILL 0x0028 /*!< Level 2 instruction cache refill */
#define ARM_PMU_L3D_CACHE_ALLOCATE 0x0029 /*!< Level 3 data cache allocation without refill */
#define ARM_PMU_L3D_CACHE_REFILL 0x002A /*!< Level 3 data cache refill */
#define ARM_PMU_L3D_CACHE 0x002B /*!< Level 3 data cache access */
#define ARM_PMU_L3D_CACHE_WB 0x002C /*!< Level 3 data cache write-back */
#define ARM_PMU_LL_CACHE_RD 0x0036 /*!< Last level data cache read */
#define ARM_PMU_LL_CACHE_MISS_RD 0x0037 /*!< Last level data cache read miss */
#define ARM_PMU_L1D_CACHE_MISS_RD 0x0039 /*!< Level 1 data cache read miss */
#define ARM_PMU_OP_COMPLETE 0x003A /*!< Operation retired */
#define ARM_PMU_OP_SPEC 0x003B /*!< Operation speculatively executed */
#define ARM_PMU_STALL 0x003C /*!< Stall cycle for instruction or operation not sent for execution */
#define ARM_PMU_STALL_OP_BACKEND 0x003D /*!< Stall cycle for instruction or operation not sent for execution due to pipeline backend */
#define ARM_PMU_STALL_OP_FRONTEND 0x003E /*!< Stall cycle for instruction or operation not sent for execution due to pipeline frontend */
#define ARM_PMU_STALL_OP 0x003F /*!< Instruction or operation slots not occupied each cycle */
#define ARM_PMU_L1D_CACHE_RD 0x0040 /*!< Level 1 data cache read */
#define ARM_PMU_LE_RETIRED 0x0100 /*!< Loop end instruction executed */
#define ARM_PMU_LE_SPEC 0x0101 /*!< Loop end instruction speculatively executed */
#define ARM_PMU_BF_RETIRED 0x0104 /*!< Branch future instruction architecturally executed and condition code check pass */
#define ARM_PMU_BF_SPEC 0x0105 /*!< Branch future instruction speculatively executed and condition code check pass */
#define ARM_PMU_LE_CANCEL 0x0108 /*!< Loop end instruction not taken */
#define ARM_PMU_BF_CANCEL 0x0109 /*!< Branch future instruction not taken */
#define ARM_PMU_SE_CALL_S 0x0114 /*!< Call to secure function, resulting in Security state change */
#define ARM_PMU_SE_CALL_NS 0x0115 /*!< Call to non-secure function, resulting in Security state change */
#define ARM_PMU_DWT_CMPMATCH0 0x0118 /*!< DWT comparator 0 match */
#define ARM_PMU_DWT_CMPMATCH1 0x0119 /*!< DWT comparator 1 match */
#define ARM_PMU_DWT_CMPMATCH2 0x011A /*!< DWT comparator 2 match */
#define ARM_PMU_DWT_CMPMATCH3 0x011B /*!< DWT comparator 3 match */
#define ARM_PMU_MVE_INST_RETIRED 0x0200 /*!< MVE instruction architecturally executed */
#define ARM_PMU_MVE_INST_SPEC 0x0201 /*!< MVE instruction speculatively executed */
#define ARM_PMU_MVE_FP_RETIRED 0x0204 /*!< MVE floating-point instruction architecturally executed */
#define ARM_PMU_MVE_FP_SPEC 0x0205 /*!< MVE floating-point instruction speculatively executed */
#define ARM_PMU_MVE_FP_HP_RETIRED 0x0208 /*!< MVE half-precision floating-point instruction architecturally executed */
#define ARM_PMU_MVE_FP_HP_SPEC 0x0209 /*!< MVE half-precision floating-point instruction speculatively executed */
#define ARM_PMU_MVE_FP_SP_RETIRED 0x020C /*!< MVE single-precision floating-point instruction architecturally executed */
#define ARM_PMU_MVE_FP_SP_SPEC 0x020D /*!< MVE single-precision floating-point instruction speculatively executed */
#define ARM_PMU_MVE_FP_MAC_RETIRED 0x0214 /*!< MVE floating-point multiply or multiply-accumulate instruction architecturally executed */
#define ARM_PMU_MVE_FP_MAC_SPEC 0x0215 /*!< MVE floating-point multiply or multiply-accumulate instruction speculatively executed */
#define ARM_PMU_MVE_INT_RETIRED 0x0224 /*!< MVE integer instruction architecturally executed */
#define ARM_PMU_MVE_INT_SPEC 0x0225 /*!< MVE integer instruction speculatively executed */
#define ARM_PMU_MVE_INT_MAC_RETIRED 0x0228 /*!< MVE multiply or multiply-accumulate instruction architecturally executed */
#define ARM_PMU_MVE_INT_MAC_SPEC 0x0229 /*!< MVE multiply or multiply-accumulate instruction speculatively executed */
#define ARM_PMU_MVE_LDST_RETIRED 0x0238 /*!< MVE load or store instruction architecturally executed */
#define ARM_PMU_MVE_LDST_SPEC 0x0239 /*!< MVE load or store instruction speculatively executed */
#define ARM_PMU_MVE_LD_RETIRED 0x023C /*!< MVE load instruction architecturally executed */
#define ARM_PMU_MVE_LD_SPEC 0x023D /*!< MVE load instruction speculatively executed */
#define ARM_PMU_MVE_ST_RETIRED 0x0240 /*!< MVE store instruction architecturally executed */
#define ARM_PMU_MVE_ST_SPEC 0x0241 /*!< MVE store instruction speculatively executed */
#define ARM_PMU_MVE_LDST_CONTIG_RETIRED 0x0244 /*!< MVE contiguous load or store instruction architecturally executed */
#define ARM_PMU_MVE_LDST_CONTIG_SPEC 0x0245 /*!< MVE contiguous load or store instruction speculatively executed */
#define ARM_PMU_MVE_LD_CONTIG_RETIRED 0x0248 /*!< MVE contiguous load instruction architecturally executed */
#define ARM_PMU_MVE_LD_CONTIG_SPEC 0x0249 /*!< MVE contiguous load instruction speculatively executed */
#define ARM_PMU_MVE_ST_CONTIG_RETIRED 0x024C /*!< MVE contiguous store instruction architecturally executed */
#define ARM_PMU_MVE_ST_CONTIG_SPEC 0x024D /*!< MVE contiguous store instruction speculatively executed */
#define ARM_PMU_MVE_LDST_NONCONTIG_RETIRED 0x0250 /*!< MVE non-contiguous load or store instruction architecturally executed */
#define ARM_PMU_MVE_LDST_NONCONTIG_SPEC 0x0251 /*!< MVE non-contiguous load or store instruction speculatively executed */
#define ARM_PMU_MVE_LD_NONCONTIG_RETIRED 0x0254 /*!< MVE non-contiguous load instruction architecturally executed */
#define ARM_PMU_MVE_LD_NONCONTIG_SPEC 0x0255 /*!< MVE non-contiguous load instruction speculatively executed */
#define ARM_PMU_MVE_ST_NONCONTIG_RETIRED 0x0258 /*!< MVE non-contiguous store instruction architecturally executed */
#define ARM_PMU_MVE_ST_NONCONTIG_SPEC 0x0259 /*!< MVE non-contiguous store instruction speculatively executed */
#define ARM_PMU_MVE_LDST_MULTI_RETIRED 0x025C /*!< MVE memory instruction targeting multiple registers architecturally executed */
#define ARM_PMU_MVE_LDST_MULTI_SPEC 0x025D /*!< MVE memory instruction targeting multiple registers speculatively executed */
#define ARM_PMU_MVE_LD_MULTI_RETIRED 0x0260 /*!< MVE memory load instruction targeting multiple registers architecturally executed */
#define ARM_PMU_MVE_LD_MULTI_SPEC 0x0261 /*!< MVE memory load instruction targeting multiple registers speculatively executed */
#define ARM_PMU_MVE_ST_MULTI_RETIRED 0x0261 /*!< MVE memory store instruction targeting multiple registers architecturally executed */
#define ARM_PMU_MVE_ST_MULTI_SPEC 0x0265 /*!< MVE memory store instruction targeting multiple registers speculatively executed */
#define ARM_PMU_MVE_LDST_UNALIGNED_RETIRED 0x028C /*!< MVE unaligned memory load or store instruction architecturally executed */
#define ARM_PMU_MVE_LDST_UNALIGNED_SPEC 0x028D /*!< MVE unaligned memory load or store instruction speculatively executed */
#define ARM_PMU_MVE_LD_UNALIGNED_RETIRED 0x0290 /*!< MVE unaligned load instruction architecturally executed */
#define ARM_PMU_MVE_LD_UNALIGNED_SPEC 0x0291 /*!< MVE unaligned load instruction speculatively executed */
#define ARM_PMU_MVE_ST_UNALIGNED_RETIRED 0x0294 /*!< MVE unaligned store instruction architecturally executed */
#define ARM_PMU_MVE_ST_UNALIGNED_SPEC 0x0295 /*!< MVE unaligned store instruction speculatively executed */
#define ARM_PMU_MVE_LDST_UNALIGNED_NONCONTIG_RETIRED 0x0298 /*!< MVE unaligned noncontiguous load or store instruction architecturally executed */
#define ARM_PMU_MVE_LDST_UNALIGNED_NONCONTIG_SPEC 0x0299 /*!< MVE unaligned noncontiguous load or store instruction speculatively executed */
#define ARM_PMU_MVE_VREDUCE_RETIRED 0x02A0 /*!< MVE vector reduction instruction architecturally executed */
#define ARM_PMU_MVE_VREDUCE_SPEC 0x02A1 /*!< MVE vector reduction instruction speculatively executed */
#define ARM_PMU_MVE_VREDUCE_FP_RETIRED 0x02A4 /*!< MVE floating-point vector reduction instruction architecturally executed */
#define ARM_PMU_MVE_VREDUCE_FP_SPEC 0x02A5 /*!< MVE floating-point vector reduction instruction speculatively executed */
#define ARM_PMU_MVE_VREDUCE_INT_RETIRED 0x02A8 /*!< MVE integer vector reduction instruction architecturally executed */
#define ARM_PMU_MVE_VREDUCE_INT_SPEC 0x02A9 /*!< MVE integer vector reduction instruction speculatively executed */
#define ARM_PMU_MVE_PRED 0x02B8 /*!< Cycles where one or more predicated beats architecturally executed */
#define ARM_PMU_MVE_STALL 0x02CC /*!< Stall cycles caused by an MVE instruction */
#define ARM_PMU_MVE_STALL_RESOURCE 0x02CD /*!< Stall cycles caused by an MVE instruction because of resource conflicts */
#define ARM_PMU_MVE_STALL_RESOURCE_MEM 0x02CE /*!< Stall cycles caused by an MVE instruction because of memory resource conflicts */
#define ARM_PMU_MVE_STALL_RESOURCE_FP 0x02CF /*!< Stall cycles caused by an MVE instruction because of floating-point resource conflicts */
#define ARM_PMU_MVE_STALL_RESOURCE_INT 0x02D0 /*!< Stall cycles caused by an MVE instruction because of integer resource conflicts */
#define ARM_PMU_MVE_STALL_BREAK 0x02D3 /*!< Stall cycles caused by an MVE chain break */
#define ARM_PMU_MVE_STALL_DEPENDENCY 0x02D4 /*!< Stall cycles caused by MVE register dependency */
#define ARM_PMU_ITCM_ACCESS 0x4007 /*!< Instruction TCM access */
#define ARM_PMU_DTCM_ACCESS 0x4008 /*!< Data TCM access */
#define ARM_PMU_TRCEXTOUT0 0x4010 /*!< ETM external output 0 */
#define ARM_PMU_TRCEXTOUT1 0x4011 /*!< ETM external output 1 */
#define ARM_PMU_TRCEXTOUT2 0x4012 /*!< ETM external output 2 */
#define ARM_PMU_TRCEXTOUT3 0x4013 /*!< ETM external output 3 */
#define ARM_PMU_CTI_TRIGOUT4 0x4018 /*!< Cross-trigger Interface output trigger 4 */
#define ARM_PMU_CTI_TRIGOUT5 0x4019 /*!< Cross-trigger Interface output trigger 5 */
#define ARM_PMU_CTI_TRIGOUT6 0x401A /*!< Cross-trigger Interface output trigger 6 */
#define ARM_PMU_CTI_TRIGOUT7 0x401B /*!< Cross-trigger Interface output trigger 7 */
/** \brief PMU Functions */
__STATIC_INLINE void ARM_PMU_Enable(void);
__STATIC_INLINE void ARM_PMU_Disable(void);
__STATIC_INLINE void ARM_PMU_Set_EVTYPER(uint32_t num, uint32_t type);
__STATIC_INLINE void ARM_PMU_CYCCNT_Reset(void);
__STATIC_INLINE void ARM_PMU_EVCNTR_ALL_Reset(void);
__STATIC_INLINE void ARM_PMU_CNTR_Enable(uint32_t mask);
__STATIC_INLINE void ARM_PMU_CNTR_Disable(uint32_t mask);
__STATIC_INLINE uint32_t ARM_PMU_Get_CCNTR(void);
__STATIC_INLINE uint32_t ARM_PMU_Get_EVCNTR(uint32_t num);
__STATIC_INLINE uint32_t ARM_PMU_Get_CNTR_OVS(void);
__STATIC_INLINE void ARM_PMU_Set_CNTR_OVS(uint32_t mask);
__STATIC_INLINE void ARM_PMU_Set_CNTR_IRQ_Enable(uint32_t mask);
__STATIC_INLINE void ARM_PMU_Set_CNTR_IRQ_Disable(uint32_t mask);
__STATIC_INLINE void ARM_PMU_CNTR_Increment(uint32_t mask);
/**
\brief Enable the PMU
*/
__STATIC_INLINE void ARM_PMU_Enable(void)
{
PMU->CTRL |= PMU_CTRL_ENABLE_Msk;
}
/**
\brief Disable the PMU
*/
__STATIC_INLINE void ARM_PMU_Disable(void)
{
PMU->CTRL &= ~PMU_CTRL_ENABLE_Msk;
}
/**
\brief Set event to count for PMU eventer counter
\param [in] num Event counter (0-30) to configure
\param [in] type Event to count
*/
__STATIC_INLINE void ARM_PMU_Set_EVTYPER(uint32_t num, uint32_t type)
{
PMU->EVTYPER[num] = type;
}
/**
\brief Reset cycle counter
*/
__STATIC_INLINE void ARM_PMU_CYCCNT_Reset(void)
{
PMU->CTRL |= PMU_CTRL_CYCCNT_RESET_Msk;
}
/**
\brief Reset all event counters
*/
__STATIC_INLINE void ARM_PMU_EVCNTR_ALL_Reset(void)
{
PMU->CTRL |= PMU_CTRL_EVENTCNT_RESET_Msk;
}
/**
\brief Enable counters
\param [in] mask Counters to enable
\note Enables one or more of the following:
- event counters (0-30)
- cycle counter
*/
__STATIC_INLINE void ARM_PMU_CNTR_Enable(uint32_t mask)
{
PMU->CNTENSET = mask;
}
/**
\brief Disable counters
\param [in] mask Counters to enable
\note Disables one or more of the following:
- event counters (0-30)
- cycle counter
*/
__STATIC_INLINE void ARM_PMU_CNTR_Disable(uint32_t mask)
{
PMU->CNTENCLR = mask;
}
/**
\brief Read cycle counter
\return Cycle count
*/
__STATIC_INLINE uint32_t ARM_PMU_Get_CCNTR(void)
{
return PMU->CCNTR;
}
/**
\brief Read event counter
\param [in] num Event counter (0-30) to read
\return Event count
*/
__STATIC_INLINE uint32_t ARM_PMU_Get_EVCNTR(uint32_t num)
{
return PMU->EVCNTR[num];
}
/**
\brief Read counter overflow status
\return Counter overflow status bits for the following:
- event counters (0-30)
- cycle counter
*/
__STATIC_INLINE uint32_t ARM_PMU_Get_CNTR_OVS(void)
{
return PMU->OVSSET;
}
/**
\brief Clear counter overflow status
\param [in] mask Counter overflow status bits to clear
\note Clears overflow status bits for one or more of the following:
- event counters (0-30)
- cycle counter
*/
__STATIC_INLINE void ARM_PMU_Set_CNTR_OVS(uint32_t mask)
{
PMU->OVSCLR = mask;
}
/**
\brief Enable counter overflow interrupt request
\param [in] mask Counter overflow interrupt request bits to set
\note Sets overflow interrupt request bits for one or more of the following:
- event counters (0-30)
- cycle counter
*/
__STATIC_INLINE void ARM_PMU_Set_CNTR_IRQ_Enable(uint32_t mask)
{
PMU->INTENSET = mask;
}
/**
\brief Disable counter overflow interrupt request
\param [in] mask Counter overflow interrupt request bits to clear
\note Clears overflow interrupt request bits for one or more of the following:
- event counters (0-30)
- cycle counter
*/
__STATIC_INLINE void ARM_PMU_Set_CNTR_IRQ_Disable(uint32_t mask)
{
PMU->INTENCLR = mask;
}
/**
\brief Software increment event counter
\param [in] mask Counters to increment
\note Software increment bits for one or more event counters (0-30)
*/
__STATIC_INLINE void ARM_PMU_CNTR_Increment(uint32_t mask)
{
PMU->SWINC = mask;
}
#endif
/******************************************************************************
* @file pmu_armv8.h
* @brief CMSIS PMU API for Armv8.1-M PMU
* @version V1.0.0
* @date 24. March 2020
******************************************************************************/
/*
* Copyright (c) 2020 Arm Limited. All rights reserved.
*
* SPDX-License-Identifier: Apache-2.0
*
* 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
*
* 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.
*/
#if defined ( __ICCARM__ )
#pragma system_include /* treat file as system include file for MISRA check */
#elif defined (__clang__)
#pragma clang system_header /* treat file as system include file */
#endif
#ifndef ARM_PMU_ARMV8_H
#define ARM_PMU_ARMV8_H
/**
* \brief PMU Events
* \note See the Armv8.1-M Architecture Reference Manual for full details on these PMU events.
* */
#define ARM_PMU_SW_INCR 0x0000 /*!< Software update to the PMU_SWINC register, architecturally executed and condition code check pass */
#define ARM_PMU_L1I_CACHE_REFILL 0x0001 /*!< L1 I-Cache refill */
#define ARM_PMU_L1D_CACHE_REFILL 0x0003 /*!< L1 D-Cache refill */
#define ARM_PMU_L1D_CACHE 0x0004 /*!< L1 D-Cache access */
#define ARM_PMU_LD_RETIRED 0x0006 /*!< Memory-reading instruction architecturally executed and condition code check pass */
#define ARM_PMU_ST_RETIRED 0x0007 /*!< Memory-writing instruction architecturally executed and condition code check pass */
#define ARM_PMU_INST_RETIRED 0x0008 /*!< Instruction architecturally executed */
#define ARM_PMU_EXC_TAKEN 0x0009 /*!< Exception entry */
#define ARM_PMU_EXC_RETURN 0x000A /*!< Exception return instruction architecturally executed and the condition code check pass */
#define ARM_PMU_PC_WRITE_RETIRED 0x000C /*!< Software change to the Program Counter (PC). Instruction is architecturally executed and condition code check pass */
#define ARM_PMU_BR_IMMED_RETIRED 0x000D /*!< Immediate branch architecturally executed */
#define ARM_PMU_BR_RETURN_RETIRED 0x000E /*!< Function return instruction architecturally executed and the condition code check pass */
#define ARM_PMU_UNALIGNED_LDST_RETIRED 0x000F /*!< Unaligned memory memory-reading or memory-writing instruction architecturally executed and condition code check pass */
#define ARM_PMU_BR_MIS_PRED 0x0010 /*!< Mispredicted or not predicted branch speculatively executed */
#define ARM_PMU_CPU_CYCLES 0x0011 /*!< Cycle */
#define ARM_PMU_BR_PRED 0x0012 /*!< Predictable branch speculatively executed */
#define ARM_PMU_MEM_ACCESS 0x0013 /*!< Data memory access */
#define ARM_PMU_L1I_CACHE 0x0014 /*!< Level 1 instruction cache access */
#define ARM_PMU_L1D_CACHE_WB 0x0015 /*!< Level 1 data cache write-back */
#define ARM_PMU_L2D_CACHE 0x0016 /*!< Level 2 data cache access */
#define ARM_PMU_L2D_CACHE_REFILL 0x0017 /*!< Level 2 data cache refill */
#define ARM_PMU_L2D_CACHE_WB 0x0018 /*!< Level 2 data cache write-back */
#define ARM_PMU_BUS_ACCESS 0x0019 /*!< Bus access */
#define ARM_PMU_MEMORY_ERROR 0x001A /*!< Local memory error */
#define ARM_PMU_INST_SPEC 0x001B /*!< Instruction speculatively executed */
#define ARM_PMU_BUS_CYCLES 0x001D /*!< Bus cycles */
#define ARM_PMU_CHAIN 0x001E /*!< For an odd numbered counter, increment when an overflow occurs on the preceding even-numbered counter on the same PE */
#define ARM_PMU_L1D_CACHE_ALLOCATE 0x001F /*!< Level 1 data cache allocation without refill */
#define ARM_PMU_L2D_CACHE_ALLOCATE 0x0020 /*!< Level 2 data cache allocation without refill */
#define ARM_PMU_BR_RETIRED 0x0021 /*!< Branch instruction architecturally executed */
#define ARM_PMU_BR_MIS_PRED_RETIRED 0x0022 /*!< Mispredicted branch instruction architecturally executed */
#define ARM_PMU_STALL_FRONTEND 0x0023 /*!< No operation issued because of the frontend */
#define ARM_PMU_STALL_BACKEND 0x0024 /*!< No operation issued because of the backend */
#define ARM_PMU_L2I_CACHE 0x0027 /*!< Level 2 instruction cache access */
#define ARM_PMU_L2I_CACHE_REFILL 0x0028 /*!< Level 2 instruction cache refill */
#define ARM_PMU_L3D_CACHE_ALLOCATE 0x0029 /*!< Level 3 data cache allocation without refill */
#define ARM_PMU_L3D_CACHE_REFILL 0x002A /*!< Level 3 data cache refill */
#define ARM_PMU_L3D_CACHE 0x002B /*!< Level 3 data cache access */
#define ARM_PMU_L3D_CACHE_WB 0x002C /*!< Level 3 data cache write-back */
#define ARM_PMU_LL_CACHE_RD 0x0036 /*!< Last level data cache read */
#define ARM_PMU_LL_CACHE_MISS_RD 0x0037 /*!< Last level data cache read miss */
#define ARM_PMU_L1D_CACHE_MISS_RD 0x0039 /*!< Level 1 data cache read miss */
#define ARM_PMU_OP_COMPLETE 0x003A /*!< Operation retired */
#define ARM_PMU_OP_SPEC 0x003B /*!< Operation speculatively executed */
#define ARM_PMU_STALL 0x003C /*!< Stall cycle for instruction or operation not sent for execution */
#define ARM_PMU_STALL_OP_BACKEND 0x003D /*!< Stall cycle for instruction or operation not sent for execution due to pipeline backend */
#define ARM_PMU_STALL_OP_FRONTEND 0x003E /*!< Stall cycle for instruction or operation not sent for execution due to pipeline frontend */
#define ARM_PMU_STALL_OP 0x003F /*!< Instruction or operation slots not occupied each cycle */
#define ARM_PMU_L1D_CACHE_RD 0x0040 /*!< Level 1 data cache read */
#define ARM_PMU_LE_RETIRED 0x0100 /*!< Loop end instruction executed */
#define ARM_PMU_LE_SPEC 0x0101 /*!< Loop end instruction speculatively executed */
#define ARM_PMU_BF_RETIRED 0x0104 /*!< Branch future instruction architecturally executed and condition code check pass */
#define ARM_PMU_BF_SPEC 0x0105 /*!< Branch future instruction speculatively executed and condition code check pass */
#define ARM_PMU_LE_CANCEL 0x0108 /*!< Loop end instruction not taken */
#define ARM_PMU_BF_CANCEL 0x0109 /*!< Branch future instruction not taken */
#define ARM_PMU_SE_CALL_S 0x0114 /*!< Call to secure function, resulting in Security state change */
#define ARM_PMU_SE_CALL_NS 0x0115 /*!< Call to non-secure function, resulting in Security state change */
#define ARM_PMU_DWT_CMPMATCH0 0x0118 /*!< DWT comparator 0 match */
#define ARM_PMU_DWT_CMPMATCH1 0x0119 /*!< DWT comparator 1 match */
#define ARM_PMU_DWT_CMPMATCH2 0x011A /*!< DWT comparator 2 match */
#define ARM_PMU_DWT_CMPMATCH3 0x011B /*!< DWT comparator 3 match */
#define ARM_PMU_MVE_INST_RETIRED 0x0200 /*!< MVE instruction architecturally executed */
#define ARM_PMU_MVE_INST_SPEC 0x0201 /*!< MVE instruction speculatively executed */
#define ARM_PMU_MVE_FP_RETIRED 0x0204 /*!< MVE floating-point instruction architecturally executed */
#define ARM_PMU_MVE_FP_SPEC 0x0205 /*!< MVE floating-point instruction speculatively executed */
#define ARM_PMU_MVE_FP_HP_RETIRED 0x0208 /*!< MVE half-precision floating-point instruction architecturally executed */
#define ARM_PMU_MVE_FP_HP_SPEC 0x0209 /*!< MVE half-precision floating-point instruction speculatively executed */
#define ARM_PMU_MVE_FP_SP_RETIRED 0x020C /*!< MVE single-precision floating-point instruction architecturally executed */
#define ARM_PMU_MVE_FP_SP_SPEC 0x020D /*!< MVE single-precision floating-point instruction speculatively executed */
#define ARM_PMU_MVE_FP_MAC_RETIRED 0x0214 /*!< MVE floating-point multiply or multiply-accumulate instruction architecturally executed */
#define ARM_PMU_MVE_FP_MAC_SPEC 0x0215 /*!< MVE floating-point multiply or multiply-accumulate instruction speculatively executed */
#define ARM_PMU_MVE_INT_RETIRED 0x0224 /*!< MVE integer instruction architecturally executed */
#define ARM_PMU_MVE_INT_SPEC 0x0225 /*!< MVE integer instruction speculatively executed */
#define ARM_PMU_MVE_INT_MAC_RETIRED 0x0228 /*!< MVE multiply or multiply-accumulate instruction architecturally executed */
#define ARM_PMU_MVE_INT_MAC_SPEC 0x0229 /*!< MVE multiply or multiply-accumulate instruction speculatively executed */
#define ARM_PMU_MVE_LDST_RETIRED 0x0238 /*!< MVE load or store instruction architecturally executed */
#define ARM_PMU_MVE_LDST_SPEC 0x0239 /*!< MVE load or store instruction speculatively executed */
#define ARM_PMU_MVE_LD_RETIRED 0x023C /*!< MVE load instruction architecturally executed */
#define ARM_PMU_MVE_LD_SPEC 0x023D /*!< MVE load instruction speculatively executed */
#define ARM_PMU_MVE_ST_RETIRED 0x0240 /*!< MVE store instruction architecturally executed */
#define ARM_PMU_MVE_ST_SPEC 0x0241 /*!< MVE store instruction speculatively executed */
#define ARM_PMU_MVE_LDST_CONTIG_RETIRED 0x0244 /*!< MVE contiguous load or store instruction architecturally executed */
#define ARM_PMU_MVE_LDST_CONTIG_SPEC 0x0245 /*!< MVE contiguous load or store instruction speculatively executed */
#define ARM_PMU_MVE_LD_CONTIG_RETIRED 0x0248 /*!< MVE contiguous load instruction architecturally executed */
#define ARM_PMU_MVE_LD_CONTIG_SPEC 0x0249 /*!< MVE contiguous load instruction speculatively executed */
#define ARM_PMU_MVE_ST_CONTIG_RETIRED 0x024C /*!< MVE contiguous store instruction architecturally executed */
#define ARM_PMU_MVE_ST_CONTIG_SPEC 0x024D /*!< MVE contiguous store instruction speculatively executed */
#define ARM_PMU_MVE_LDST_NONCONTIG_RETIRED 0x0250 /*!< MVE non-contiguous load or store instruction architecturally executed */
#define ARM_PMU_MVE_LDST_NONCONTIG_SPEC 0x0251 /*!< MVE non-contiguous load or store instruction speculatively executed */
#define ARM_PMU_MVE_LD_NONCONTIG_RETIRED 0x0254 /*!< MVE non-contiguous load instruction architecturally executed */
#define ARM_PMU_MVE_LD_NONCONTIG_SPEC 0x0255 /*!< MVE non-contiguous load instruction speculatively executed */
#define ARM_PMU_MVE_ST_NONCONTIG_RETIRED 0x0258 /*!< MVE non-contiguous store instruction architecturally executed */
#define ARM_PMU_MVE_ST_NONCONTIG_SPEC 0x0259 /*!< MVE non-contiguous store instruction speculatively executed */
#define ARM_PMU_MVE_LDST_MULTI_RETIRED 0x025C /*!< MVE memory instruction targeting multiple registers architecturally executed */
#define ARM_PMU_MVE_LDST_MULTI_SPEC 0x025D /*!< MVE memory instruction targeting multiple registers speculatively executed */
#define ARM_PMU_MVE_LD_MULTI_RETIRED 0x0260 /*!< MVE memory load instruction targeting multiple registers architecturally executed */
#define ARM_PMU_MVE_LD_MULTI_SPEC 0x0261 /*!< MVE memory load instruction targeting multiple registers speculatively executed */
#define ARM_PMU_MVE_ST_MULTI_RETIRED 0x0261 /*!< MVE memory store instruction targeting multiple registers architecturally executed */
#define ARM_PMU_MVE_ST_MULTI_SPEC 0x0265 /*!< MVE memory store instruction targeting multiple registers speculatively executed */
#define ARM_PMU_MVE_LDST_UNALIGNED_RETIRED 0x028C /*!< MVE unaligned memory load or store instruction architecturally executed */
#define ARM_PMU_MVE_LDST_UNALIGNED_SPEC 0x028D /*!< MVE unaligned memory load or store instruction speculatively executed */
#define ARM_PMU_MVE_LD_UNALIGNED_RETIRED 0x0290 /*!< MVE unaligned load instruction architecturally executed */
#define ARM_PMU_MVE_LD_UNALIGNED_SPEC 0x0291 /*!< MVE unaligned load instruction speculatively executed */
#define ARM_PMU_MVE_ST_UNALIGNED_RETIRED 0x0294 /*!< MVE unaligned store instruction architecturally executed */
#define ARM_PMU_MVE_ST_UNALIGNED_SPEC 0x0295 /*!< MVE unaligned store instruction speculatively executed */
#define ARM_PMU_MVE_LDST_UNALIGNED_NONCONTIG_RETIRED 0x0298 /*!< MVE unaligned noncontiguous load or store instruction architecturally executed */
#define ARM_PMU_MVE_LDST_UNALIGNED_NONCONTIG_SPEC 0x0299 /*!< MVE unaligned noncontiguous load or store instruction speculatively executed */
#define ARM_PMU_MVE_VREDUCE_RETIRED 0x02A0 /*!< MVE vector reduction instruction architecturally executed */
#define ARM_PMU_MVE_VREDUCE_SPEC 0x02A1 /*!< MVE vector reduction instruction speculatively executed */
#define ARM_PMU_MVE_VREDUCE_FP_RETIRED 0x02A4 /*!< MVE floating-point vector reduction instruction architecturally executed */
#define ARM_PMU_MVE_VREDUCE_FP_SPEC 0x02A5 /*!< MVE floating-point vector reduction instruction speculatively executed */
#define ARM_PMU_MVE_VREDUCE_INT_RETIRED 0x02A8 /*!< MVE integer vector reduction instruction architecturally executed */
#define ARM_PMU_MVE_VREDUCE_INT_SPEC 0x02A9 /*!< MVE integer vector reduction instruction speculatively executed */
#define ARM_PMU_MVE_PRED 0x02B8 /*!< Cycles where one or more predicated beats architecturally executed */
#define ARM_PMU_MVE_STALL 0x02CC /*!< Stall cycles caused by an MVE instruction */
#define ARM_PMU_MVE_STALL_RESOURCE 0x02CD /*!< Stall cycles caused by an MVE instruction because of resource conflicts */
#define ARM_PMU_MVE_STALL_RESOURCE_MEM 0x02CE /*!< Stall cycles caused by an MVE instruction because of memory resource conflicts */
#define ARM_PMU_MVE_STALL_RESOURCE_FP 0x02CF /*!< Stall cycles caused by an MVE instruction because of floating-point resource conflicts */
#define ARM_PMU_MVE_STALL_RESOURCE_INT 0x02D0 /*!< Stall cycles caused by an MVE instruction because of integer resource conflicts */
#define ARM_PMU_MVE_STALL_BREAK 0x02D3 /*!< Stall cycles caused by an MVE chain break */
#define ARM_PMU_MVE_STALL_DEPENDENCY 0x02D4 /*!< Stall cycles caused by MVE register dependency */
#define ARM_PMU_ITCM_ACCESS 0x4007 /*!< Instruction TCM access */
#define ARM_PMU_DTCM_ACCESS 0x4008 /*!< Data TCM access */
#define ARM_PMU_TRCEXTOUT0 0x4010 /*!< ETM external output 0 */
#define ARM_PMU_TRCEXTOUT1 0x4011 /*!< ETM external output 1 */
#define ARM_PMU_TRCEXTOUT2 0x4012 /*!< ETM external output 2 */
#define ARM_PMU_TRCEXTOUT3 0x4013 /*!< ETM external output 3 */
#define ARM_PMU_CTI_TRIGOUT4 0x4018 /*!< Cross-trigger Interface output trigger 4 */
#define ARM_PMU_CTI_TRIGOUT5 0x4019 /*!< Cross-trigger Interface output trigger 5 */
#define ARM_PMU_CTI_TRIGOUT6 0x401A /*!< Cross-trigger Interface output trigger 6 */
#define ARM_PMU_CTI_TRIGOUT7 0x401B /*!< Cross-trigger Interface output trigger 7 */
/** \brief PMU Functions */
__STATIC_INLINE void ARM_PMU_Enable(void);
__STATIC_INLINE void ARM_PMU_Disable(void);
__STATIC_INLINE void ARM_PMU_Set_EVTYPER(uint32_t num, uint32_t type);
__STATIC_INLINE void ARM_PMU_CYCCNT_Reset(void);
__STATIC_INLINE void ARM_PMU_EVCNTR_ALL_Reset(void);
__STATIC_INLINE void ARM_PMU_CNTR_Enable(uint32_t mask);
__STATIC_INLINE void ARM_PMU_CNTR_Disable(uint32_t mask);
__STATIC_INLINE uint32_t ARM_PMU_Get_CCNTR(void);
__STATIC_INLINE uint32_t ARM_PMU_Get_EVCNTR(uint32_t num);
__STATIC_INLINE uint32_t ARM_PMU_Get_CNTR_OVS(void);
__STATIC_INLINE void ARM_PMU_Set_CNTR_OVS(uint32_t mask);
__STATIC_INLINE void ARM_PMU_Set_CNTR_IRQ_Enable(uint32_t mask);
__STATIC_INLINE void ARM_PMU_Set_CNTR_IRQ_Disable(uint32_t mask);
__STATIC_INLINE void ARM_PMU_CNTR_Increment(uint32_t mask);
/**
\brief Enable the PMU
*/
__STATIC_INLINE void ARM_PMU_Enable(void)
{
PMU->CTRL |= PMU_CTRL_ENABLE_Msk;
}
/**
\brief Disable the PMU
*/
__STATIC_INLINE void ARM_PMU_Disable(void)
{
PMU->CTRL &= ~PMU_CTRL_ENABLE_Msk;
}
/**
\brief Set event to count for PMU eventer counter
\param [in] num Event counter (0-30) to configure
\param [in] type Event to count
*/
__STATIC_INLINE void ARM_PMU_Set_EVTYPER(uint32_t num, uint32_t type)
{
PMU->EVTYPER[num] = type;
}
/**
\brief Reset cycle counter
*/
__STATIC_INLINE void ARM_PMU_CYCCNT_Reset(void)
{
PMU->CTRL |= PMU_CTRL_CYCCNT_RESET_Msk;
}
/**
\brief Reset all event counters
*/
__STATIC_INLINE void ARM_PMU_EVCNTR_ALL_Reset(void)
{
PMU->CTRL |= PMU_CTRL_EVENTCNT_RESET_Msk;
}
/**
\brief Enable counters
\param [in] mask Counters to enable
\note Enables one or more of the following:
- event counters (0-30)
- cycle counter
*/
__STATIC_INLINE void ARM_PMU_CNTR_Enable(uint32_t mask)
{
PMU->CNTENSET = mask;
}
/**
\brief Disable counters
\param [in] mask Counters to enable
\note Disables one or more of the following:
- event counters (0-30)
- cycle counter
*/
__STATIC_INLINE void ARM_PMU_CNTR_Disable(uint32_t mask)
{
PMU->CNTENCLR = mask;
}
/**
\brief Read cycle counter
\return Cycle count
*/
__STATIC_INLINE uint32_t ARM_PMU_Get_CCNTR(void)
{
return PMU->CCNTR;
}
/**
\brief Read event counter
\param [in] num Event counter (0-30) to read
\return Event count
*/
__STATIC_INLINE uint32_t ARM_PMU_Get_EVCNTR(uint32_t num)
{
return PMU->EVCNTR[num];
}
/**
\brief Read counter overflow status
\return Counter overflow status bits for the following:
- event counters (0-30)
- cycle counter
*/
__STATIC_INLINE uint32_t ARM_PMU_Get_CNTR_OVS(void)
{
return PMU->OVSSET;
}
/**
\brief Clear counter overflow status
\param [in] mask Counter overflow status bits to clear
\note Clears overflow status bits for one or more of the following:
- event counters (0-30)
- cycle counter
*/
__STATIC_INLINE void ARM_PMU_Set_CNTR_OVS(uint32_t mask)
{
PMU->OVSCLR = mask;
}
/**
\brief Enable counter overflow interrupt request
\param [in] mask Counter overflow interrupt request bits to set
\note Sets overflow interrupt request bits for one or more of the following:
- event counters (0-30)
- cycle counter
*/
__STATIC_INLINE void ARM_PMU_Set_CNTR_IRQ_Enable(uint32_t mask)
{
PMU->INTENSET = mask;
}
/**
\brief Disable counter overflow interrupt request
\param [in] mask Counter overflow interrupt request bits to clear
\note Clears overflow interrupt request bits for one or more of the following:
- event counters (0-30)
- cycle counter
*/
__STATIC_INLINE void ARM_PMU_Set_CNTR_IRQ_Disable(uint32_t mask)
{
PMU->INTENCLR = mask;
}
/**
\brief Software increment event counter
\param [in] mask Counters to increment
\note Software increment bits for one or more event counters (0-30)
*/
__STATIC_INLINE void ARM_PMU_CNTR_Increment(uint32_t mask)
{
PMU->SWINC = mask;
}
#endif

15
libraries/cmsis/cm4/device_support/at32f403a_407.h
View File

@@ -1,8 +1,8 @@
/**
**************************************************************************
* @file at32f403a_407.h
* @version v2.0.4
* @date 2021-11-26
* @version v2.0.6
* @date 2021-12-31
* @brief at32f403a_407 header file
**************************************************************************
* Copyright notice & Disclaimer
@@ -60,7 +60,7 @@ extern "C" {
!defined (AT32F407VCT7) && !defined (AT32F407RET7) && !defined (AT32F407VET7) && \
!defined (AT32F407AVCT7) && !defined (AT32F407AVGT7)
#error "Please select first the target at32f4xx device used in your application (in at32f4xx.h file)"
#error "Please select first the target device used in your application (in at32f403a_407.h file)"
#endif
#if defined (AT32F403AVCT7) || defined (AT32F403ARCT7) || defined (AT32F403ACCT7) || \
@@ -97,7 +97,7 @@ extern "C" {
*/
#define __AT32F403A_407_LIBRARY_VERSION_MAJOR (0x02) /*!< [31:24] major version */
#define __AT32F403A_407_LIBRARY_VERSION_MIDDLE (0x00) /*!< [23:16] middle version */
#define __AT32F403A_407_LIBRARY_VERSION_MINOR (0x04) /*!< [15:8] minor version */
#define __AT32F403A_407_LIBRARY_VERSION_MINOR (0x06) /*!< [15:8] minor version */
#define __AT32F403A_407_LIBRARY_VERSION_RC (0x00) /*!< [7:0] release candidate */
#define __AT32F403A_407_LIBRARY_VERSION ((__AT32F403A_407_LIBRARY_VERSION_MAJOR << 24) | \
(__AT32F403A_407_LIBRARY_VERSION_MIDDLE << 16) | \
@@ -548,11 +548,7 @@ typedef enum {ERROR = 0, SUCCESS = !ERROR} error_status;
#define SDIO2_BASE (AHBPERIPH_BASE + 0x3400)
#define EMAC_BASE (AHBPERIPH_BASE + 0x8000)
#define XMC_BANK1_REG_BASE (XMC_REG_BASE + 0x0000)
#define XMC_BANK1E_REG_BASE (XMC_REG_BASE + 0x0104)
#define XMC_BANK1E_H_BASE (XMC_REG_BASE + 0x0220)
#define XMC_BANK2_REG_BASE (XMC_REG_BASE + 0x0060)
#define XMC_BANK3_REG_BASE (XMC_REG_BASE + 0x0080)
#define XMC_BANK4_REG_BASE (XMC_REG_BASE + 0x00A0)
#define EMAC_MMC_BASE (EMAC_BASE + 0x0100)
#define EMAC_PTP_BASE (EMAC_BASE + 0x0700)
#define EMAC_DMA_BASE (EMAC_BASE + 0x1000)
@@ -569,7 +565,8 @@ typedef enum {ERROR = 0, SUCCESS = !ERROR} error_status;
/**
* @}
*/
#include "at32f403a_407_def.h"
#include "at32f403a_407_conf.h"
#ifdef __cplusplus

4
libraries/cmsis/cm4/device_support/at32f403a_407_conf_template.h
View File

@@ -1,8 +1,8 @@
/**
**************************************************************************
* @file at32f403a_407_conf.h
* @version v2.0.4
* @date 2021-11-26
* @version v2.0.6
* @date 2021-12-31
* @brief at32f403a_407 config header file
**************************************************************************
* Copyright notice & Disclaimer

4
libraries/cmsis/cm4/device_support/startup/gcc/startup_at32f403a_407.s
View File

@@ -1,8 +1,8 @@
/**
******************************************************************************
* @file startup_at32f403a_407.s
* @version v2.0.4
* @date 2021-11-26
* @version v2.0.6
* @date 2021-12-31
* @brief at32f403a_407xx devices vector table for gcc toolchain.
* this module performs:
* - set the initial sp

4
libraries/cmsis/cm4/device_support/startup/iar/startup_at32f403a_407.s
View File

@@ -1,7 +1,7 @@
;**************************************************************************
;* @file startup_at32f403a_407.s
;* @version v2.0.4
;* @date 2021-11-26
;* @version v2.0.6
;* @date 2021-12-31
;* @brief at32f403a_407 startup file for IAR Systems
;**************************************************************************
;

4
libraries/cmsis/cm4/device_support/startup/mdk/startup_at32f403a_407.s
View File

@@ -1,7 +1,7 @@
;**************************************************************************
;* @file startup_at32f403a_407.s
;* @version v2.0.4
;* @date 2021-11-26
;* @version v2.0.6
;* @date 2021-12-31
;* @brief at32f403a_407 startup file for keil
;**************************************************************************
;

4
libraries/cmsis/cm4/device_support/system_at32f403a_407.c
View File

@@ -1,8 +1,8 @@
/**
**************************************************************************
* @file system_at32f403a_407.c
* @version v2.0.4
* @date 2021-11-26
* @version v2.0.6
* @date 2021-12-31
* @brief contains all the functions for cmsis cortex-m4 system source file
**************************************************************************
* Copyright notice & Disclaimer

4
libraries/cmsis/cm4/device_support/system_at32f403a_407.h
View File

@@ -1,8 +1,8 @@
/**
**************************************************************************
* @file system_at32f403a_407.h
* @version v2.0.4
* @date 2021-11-26
* @version v2.0.6
* @date 2021-12-31
* @brief cmsis cortex-m4 system header file.
**************************************************************************
* Copyright notice & Disclaimer