目录

0.写在前面

1.CLA介绍

2.CLA使用前提

3. 初始化程序

4.内存管理

4.1 RAM运行

4.2 FLASH运行

5. 运行程序

5.库文件

6.最后移植完成后的文件

7. 最后效果

0.写在前面

CLA的整个移植过程非常复杂，如果对于初学者可以从下往上看，先能移植起来再看原理

好多人都在要这份工程，我的工程都开源在gitee上了，链接在下面，如果需要可以自取。如果链接失效等问题可以评论区里告诉我，加油。

TMS320F28377D_AloneCPU_CodeLearn: TMS320F2877D单核程序开发 (gitee.com)

1.CLA介绍

CLA是什么呢？网上有很多说法，什么什么规律加速器，一堆名词着实头疼。其实很简单，在使用DSP的时候，总会有大量的计算，从简单的加减乘除，到正弦，到快速傅里叶，滤波器等都会要计算。此时，如果用CPU就不能操控其他外设了，在这种情况下就可以用CLA了。CLA你可以理解成DSP中专门用于计算的模块，正因如此，CLA不能初始化模块，配置模块，相反在计算是CLA的专场。如果读者用过cortex-M4内核的芯片就知道M4内核的芯片有对DSP库的支持，在DSP里CLA就如同MCU中DSP的作用，是专门用于计算的模块。

2.CLA使用前提

首先在开始前你要对TI的CMD的文件，如果读者是首次接触或是像我一样是STM32开发者转DSP的话，要有一个准备去写CMD文件，那么什么是CMD文件呢？假设芯片内的存储是一个仓库，那么你就是仓库管理员，你需要对仓库的每一部分都很熟悉，哪些地方是可以堆放“危险品”的，哪些地方可以堆放“长期滞留”品的，在STM32中几乎很少会考虑到这些内容，但是在使用DSP的东西是难以避免的。

705c24de1d1e435eae1b6cec66f6ab8d.png

上图是F28377D的内部存储器映射

可以看到有些区域是CLA可以访问，有些区域是DMA访问的所以在这种情况下你要对这些内存区域熟悉，并能合理的管理每一块区域。通过编写CMD文件合理地管理每一块区域。在CLA的使用中随时都需要你关注内存的情况。

3. 初始化程序

/*
 * CLA_init.c
 *
 *  Created on: 2023年5月1日
 *      Author: 18752
 */

#include "CLA_init.h"



#pragma DATA_SECTION(fVal,"CpuToCla1MsgRAM");
float fVal;
#pragma DATA_SECTION(fResult,"Cla1ToCpuMsgRAM");
float fResult;

#pragma DATA_SECTION(IOBuffer,"IOBuffer")
float IOBuffer[(256+1)*2] = {
    #include "ffttest.h"
};


//
// CLA_runTest - Execute CLA task tests for specified vectors
//
void CLA_runTest(unsigned char RTest)
{
    switch(RTest)
    {
        case 1: CLA_forceTasks(CLA1_BASE,CLA_TASKFLAG_1);   break;
        case 2: CLA_forceTasks(CLA1_BASE,CLA_TASKFLAG_2);   break;
    }
    WAITSTEP;
}

void CLA_init()
{
    EALLOW;

    CLA_DSP_initEpie();

    CLA_DSP_configClaMemory();

    CLA_DSP_initCpu1Cla1();

    EDIS;
}

void CLA_DSP_initEpie()
{
    Interrupt_initModule();
    Interrupt_initVectorTable();
}


void CLA_DSP_configClaMemory(void)
{
    extern uint32_t Cla1ProgRunStart, Cla1ProgLoadStart, Cla1ProgLoadSize;
    extern uint32_t CLA1fftTablesRunStart, CLA1fftTablesLoadStart, CLA1fftTablesLoadSize;

    EALLOW;

#ifdef _FLASH
    // Copy over code and tables from FLASH to RAM
    memcpy((uint32_t *)&Cla1ProgRunStart, (uint32_t *)&Cla1ProgLoadStart,
            (uint32_t)&Cla1ProgLoadSize);

    memcpy((uint32_t *)&CLA1fftTablesRunStart, (uint32_t *)&CLA1fftTablesLoadStart,
            (uint32_t)&CLA1fftTablesLoadSize);
#endif //_FLASH

    // Initialize and wait for CLA1ToCPUMsgRAM
    MemCfg_initSections(MEMCFG_SECT_MSGCLA1TOCPU);
    while (!MemCfg_getInitStatus(MEMCFG_SECT_MSGCLA1TOCPU)){};

    // Initialize and wait for CPUToCLA1MsgRAM
    MemCfg_initSections(MEMCFG_SECT_MSGCPUTOCLA1);
    while (!MemCfg_getInitStatus(MEMCFG_SECT_MSGCPUTOCLA1)){};

    // Select LS0-LS1RAM to be the programming space for the CLA
    // First configure the CLA to be the master for LS0 and then
    // set the space to be a program block
    MemCfg_setLSRAMMasterSel(MEMCFG_SECT_LS0,MEMCFG_LSRAMMASTER_CPU_CLA1);
    MemCfg_setCLAMemType(MEMCFG_SECT_LS0,MEMCFG_CLA_MEM_PROGRAM);
    MemCfg_setLSRAMMasterSel(MEMCFG_SECT_LS1,MEMCFG_LSRAMMASTER_CPU_CLA1);
    MemCfg_setCLAMemType(MEMCFG_SECT_LS1,MEMCFG_CLA_MEM_PROGRAM);

    //Next configure LS2-LS5RAM as data spaces for the CLA
    // First configure the CLA to be the master for LSx and then
    // set the spaces to be code blocks

    MemCfg_setLSRAMMasterSel(MEMCFG_SECT_LS2,MEMCFG_LSRAMMASTER_CPU_CLA1);
    MemCfg_setCLAMemType(MEMCFG_SECT_LS2, MEMCFG_CLA_MEM_DATA);

    MemCfg_setLSRAMMasterSel(MEMCFG_SECT_LS3,MEMCFG_LSRAMMASTER_CPU_CLA1);
    MemCfg_setCLAMemType(MEMCFG_SECT_LS3, MEMCFG_CLA_MEM_DATA);

    MemCfg_setLSRAMMasterSel(MEMCFG_SECT_LS4,MEMCFG_LSRAMMASTER_CPU_CLA1);
    MemCfg_setCLAMemType(MEMCFG_SECT_LS4, MEMCFG_CLA_MEM_DATA);

    MemCfg_setLSRAMMasterSel(MEMCFG_SECT_LS5,MEMCFG_LSRAMMASTER_CPU_CLA1);
    MemCfg_setCLAMemType(MEMCFG_SECT_LS5, MEMCFG_CLA_MEM_DATA);

}

void CLA_DSP_initCpu1Cla1(void)
{
    // Compute all CLA task vectors
    // On Type-1 CLAs the MVECT registers accept full 16-bit task addresses as
    // opposed to offsets used on older Type-0 CLAs

//
// Suppressing #770-D conversion from pointer to smaller integer
// The CLA address range is 16 bits so the addresses passed to the MVECT
// registers will be in the lower 64KW address space. Turn the warning
// back on after the MVECTs are assigned addresses
//
#pragma diag_suppress=770
    CLA_mapTaskVector(CLA1_BASE,CLA_MVECT_1,(uint16_t)&Cla1Task1);
    CLA_mapTaskVector(CLA1_BASE,CLA_MVECT_2,(uint16_t)&Cla1Task2);
    CLA_mapTaskVector(CLA1_BASE,CLA_MVECT_3,(uint16_t)&Cla1Task3);
    CLA_mapTaskVector(CLA1_BASE,CLA_MVECT_4,(uint16_t)&Cla1Task4);
    CLA_mapTaskVector(CLA1_BASE,CLA_MVECT_5,(uint16_t)&Cla1Task5);
    CLA_mapTaskVector(CLA1_BASE,CLA_MVECT_6,(uint16_t)&Cla1Task6);
    CLA_mapTaskVector(CLA1_BASE,CLA_MVECT_7,(uint16_t)&Cla1Task7);
    CLA_mapTaskVector(CLA1_BASE,CLA_MVECT_8,(uint16_t)&Cla1Task8);

#pragma diag_warning=770

    // Enable IACK instruction to start a task on CLA in software
    // for all  8 CLA tasks

    CLA_enableIACK(CLA1_BASE);
    CLA_enableTasks(CLA1_BASE, CLA_TASKFLAG_ALL);

    // Configure the vectors for the end-of-task interrupt for all
    // 8 tasks

    Interrupt_register(INT_CLA1_1, &cla1Isr1);
    Interrupt_register(INT_CLA1_2, &cla1Isr2);
    Interrupt_register(INT_CLA1_3, &cla1Isr3);
    Interrupt_register(INT_CLA1_4, &cla1Isr4);
    Interrupt_register(INT_CLA1_5, &cla1Isr5);
    Interrupt_register(INT_CLA1_6, &cla1Isr6);
    Interrupt_register(INT_CLA1_7, &cla1Isr7);
    Interrupt_register(INT_CLA1_8, &cla1Isr8);
}

这是CLA的存储分配，没错程序里也要有。这是初始化部分，它管理着你内存是如何分配的，没有这部分CLA无法完成初始化。

void CLA_DSP_initEpie()
{
    Interrupt_initModule();
    Interrupt_initVectorTable();
}


void CLA_DSP_initCpu1Cla1(void)
{
    // Compute all CLA task vectors
    // On Type-1 CLAs the MVECT registers accept full 16-bit task addresses as
    // opposed to offsets used on older Type-0 CLAs

//
// Suppressing #770-D conversion from pointer to smaller integer
// The CLA address range is 16 bits so the addresses passed to the MVECT
// registers will be in the lower 64KW address space. Turn the warning
// back on after the MVECTs are assigned addresses
//
#pragma diag_suppress=770
    CLA_mapTaskVector(CLA1_BASE,CLA_MVECT_1,(uint16_t)&Cla1Task1);
    CLA_mapTaskVector(CLA1_BASE,CLA_MVECT_2,(uint16_t)&Cla1Task2);
    CLA_mapTaskVector(CLA1_BASE,CLA_MVECT_3,(uint16_t)&Cla1Task3);
    CLA_mapTaskVector(CLA1_BASE,CLA_MVECT_4,(uint16_t)&Cla1Task4);
    CLA_mapTaskVector(CLA1_BASE,CLA_MVECT_5,(uint16_t)&Cla1Task5);
    CLA_mapTaskVector(CLA1_BASE,CLA_MVECT_6,(uint16_t)&Cla1Task6);
    CLA_mapTaskVector(CLA1_BASE,CLA_MVECT_7,(uint16_t)&Cla1Task7);
    CLA_mapTaskVector(CLA1_BASE,CLA_MVECT_8,(uint16_t)&Cla1Task8);

#pragma diag_warning=770

    // Enable IACK instruction to start a task on CLA in software
    // for all  8 CLA tasks

    CLA_enableIACK(CLA1_BASE);
    CLA_enableTasks(CLA1_BASE, CLA_TASKFLAG_ALL);

    // Configure the vectors for the end-of-task interrupt for all
    // 8 tasks

    Interrupt_register(INT_CLA1_1, &cla1Isr1);
    Interrupt_register(INT_CLA1_2, &cla1Isr2);
    Interrupt_register(INT_CLA1_3, &cla1Isr3);
    Interrupt_register(INT_CLA1_4, &cla1Isr4);
    Interrupt_register(INT_CLA1_5, &cla1Isr5);
    Interrupt_register(INT_CLA1_6, &cla1Isr6);
    Interrupt_register(INT_CLA1_7, &cla1Isr7);
    Interrupt_register(INT_CLA1_8, &cla1Isr8);
}

这是官方的程序，是对CLA中断之类的初始化

void CLA_init()
{
    EALLOW;

    CLA_DSP_initEpie();

    CLA_DSP_configClaMemory();

    CLA_DSP_initCpu1Cla1();

    EDIS;
}

总初始化

void CLA_runTest(unsigned char RTest)
{
    switch(RTest)
    {
        case 1: CLA_forceTasks(CLA1_BASE,CLA_TASKFLAG_1);   break;
        case 2: CLA_forceTasks(CLA1_BASE,CLA_TASKFLAG_2);   break;
    }
    WAITSTEP;
}

这是运行CLA程序，有点类似ADC软件触发一样

上述部分在官方的例程中都有，基本上只要无脑搬基本也没什么问题。

4.内存管理

关于CLA的内存分配问题上，我们主要关注点是LSx RAM内存区域，因为这个区域是CLA运行区域。

fc9bc54abf4d4bd08e512ce42bef3351.png

可以看到从LS0-LS5的区域CLA是可以存取的所有的CLA程序，运行中所产生的内容都会在这部分区域。除此之外还有一部分就是MSGRAM1和2就是第四行和第五行的内容，这部分内容是CPU与CLA的通信的“桥梁”，通过两个区域，CPU的数据可以转移到CLA内，同时CLA的内容也可以转移到CPU内。

4.1 RAM运行


CLA_SCRATCHPAD_SIZE = 0x100;
--undef_sym=__cla_scratchpad_end
--undef_sym=__cla_scratchpad_start

MEMORY
{
//一般用于存放代码
PAGE 0 :
	BEGIN				: origin = 0x000000, length = 0x000002

	RAMGS0      		: origin = 0x00C000, length = 0x001000
	RAMGS1      		: origin = 0x00D000, length = 0x001000
	RAMGS2      		: origin = 0x00E000, length = 0x001000
	RAMGS3      		: origin = 0x00F000, length = 0x001000
	RAMGS4      		: origin = 0x010000, length = 0x001000
	RAMGS5      		: origin = 0x011000, length = 0x001000


	RESET				: origin = 0x3FFFC0, length = 0x000002
	IQTABLES			: origin = 0x3FE000, length = 0x000B50     /* IQ Math Tables in Boot ROM */
	IQTABLES2			: origin = 0x3FEB50, length = 0x00008C
	IQTABLES3			: origin = 0x3FEBDC, length = 0x0000AA

	FLASHA				: origin = 0x080002, length = 0x001FFE   /* on-chip Flash */
	FLASHB				: origin = 0x082000, length = 0x002000   /* on-chip Flash */
	FLASHC				: origin = 0x084000, length = 0x002000   /* on-chip Flash */
	FLASHD				: origin = 0x086000, length = 0x002000   /* on-chip Flash */
	FLASHE				: origin = 0x088000, length = 0x008000   /* on-chip Flash */
	FLASHF				: origin = 0x090000, length = 0x008000   /* on-chip Flash */
	FLASHG				: origin = 0x098000, length = 0x008000   /* on-chip Flash */


//一般用于存放数据
PAGE 1 :
	BOOT_RSVD         	: origin = 0x000002, length = 0x00004E   /* Part of M0, BOOT rom will use this for stack */

	RAMM0             	: origin = 0x000050, length = 0x0003B0
	RAMM1             	: origin = 0x000400, length = 0x000400

	RAMD0            	: origin = 0x00B000, length = 0x000800
	RAMD1            	: origin = 0x00B800, length = 0x000800

	RAMGS6      		: origin = 0x012000, length = 0x001000
	RAMGS7      		: origin = 0x013000, length = 0x001000
	RAMGS8				: origin = 0x014000, length = 0x001000
	RAMGS9				: origin = 0x015000, length = 0x001000
	RAMGS10				: origin = 0x016000, length = 0x001000
	RAMGS11				: origin = 0x017000, length = 0x001000
	RAMGS12				: origin = 0x018000, length = 0x001000
	RAMGS13				: origin = 0x019000, length = 0x001000
	RAMGS14				: origin = 0x01A000, length = 0x001000
	RAMGS15				: origin = 0x01B000, length = 0x001000

	RAMLS0_1			: origin = 0x008000, length = 0x002000
	RAMLS2_3			: origin = 0x00A000, length = 0x000A00
	RAMLS4_5			: origin = 0x00AA00, length = 0x000600

	CLA1_MSGRAMLOW		: origin = 0x001480, length = 0x000080
	CLA1_MSGRAMHIGH		: origin = 0x001500, length = 0x000080

	CPU2TOCPU1RAM   	: origin = 0x03F800, length = 0x000400
	CPU1TOCPU2RAM   	: origin = 0x03FC00, length = 0x000400

	FLASHH				: origin = 0x0A0000, length = 0x008000   /* on-chip Flash */
	FLASHI				: origin = 0x0A8000, length = 0x008000   /* on-chip Flash */
	FLASHJ				: origin = 0x0B0000, length = 0x008000   /* on-chip Flash */
	FLASHK 				: origin = 0x0B8000, length = 0x002000   /* on-chip Flash */
	FLASHL				: origin = 0x0BA000, length = 0x002000   /* on-chip Flash */
	FLASHM				: origin = 0x0BC000, length = 0x002000   /* on-chip Flash */
	FLASHN				: origin = 0x0BE000, length = 0x002000   /* on-chip Flash */
}

SECTIONS
{
	codestart		: > BEGIN,      				PAGE = 0

	//可存储在flash的字段
	.text			:>> RAMGS0 | RAMGS1 | RAMGS2,	PAGE = 0	//程序段
	.cinit			: > RAMGS4,      				PAGE = 0	//初始化数据
	.const			: > RAMGS4,						PAGE = 0	//在EABI mode下与cinit一致,初始化段上电后会复制到bss
	.bss			: > RAMGS4,						PAGE = 0	//全局变量和静态变量
	.bss:output		: > RAMGS4,						PAGE = 0

	//重要字段
	ramfuncs		: > RAMM0,						PAGE = 1
	.TI.ramfunc		: > RAMM0,						PAGE = 1
	.switch			: > RAMM0,      				PAGE = 1	//switch字段
	.init_array		: > RAMM0,						PAGE = 1	//启动时调用的 C++ 构造函数的表
	.stack			: > RAMM1,      				PAGE = 1	//堆栈，最低至少需要64k
	.data			: > RAMGS7,						PAGE = 1	//程序运行中所产生的数据

	//非重要内存字段
	.econst			: > RAMGS11,					PAGE = 1
	.ebss			: > RAMGS11,					PAGE = 1	//为使用大寄存器模式时的全局变量和静态变量预留的空间，在程序上电时
																//cinit空间中的数据复制出来并存储在.ebss中
	.pinit			: > RAMGS11,					PAGE = 1	//启动时要调用的构造函数表
	.cio			: > RAMGS11,					PAGE = 1	//printf等输入输出函数使用的缓冲区所在的段
	.sysmem			: > RAMGS11,					PAGE = 1	//maclloc所用的段


	//CLA字段
	IOBuffer         : > RAMLS2_3,					PAGE = 1
	Cla1Prog         : > RAMLS0_1,					PAGE = 1
	Cla1Prog:cla_dsp_fft : > RAMLS0_1, HIGH			PAGE = 1

	ramfuncs         : LOAD = FLASHL,
                       RUN = RAMD0,
                       RUN_START(RamfuncsRunStart),
                       LOAD_START(RamfuncsLoadStart),
                       LOAD_SIZE(RamfuncsLoadSize),
                       PAGE = 1

	CLA1fftTables    : LOAD = FLASHM,
                       RUN = RAMLS4_5,
                       RUN_START(CLA1fftTablesRunStart),
                       LOAD_START(CLA1fftTablesLoadStart),
                       LOAD_SIZE(CLA1fftTablesLoadSize),
                       PAGE = 1

	Cla1Prog         : LOAD = FLASHM,
                       RUN = RAMLS0_1,
                       RUN_START(Cla1ProgRunStart),
                       LOAD_START(Cla1ProgLoadStart),
                       LOAD_SIZE(Cla1ProgLoadSize),
                       PAGE = 1


	Cla1ToCpuMsgRAM  : > CLA1_MSGRAMLOW,   PAGE = 1
	CpuToCla1MsgRAM  : > CLA1_MSGRAMHIGH,  PAGE = 1

	/* CLA C compiler sections */
	//
	// Must be allocated to memory the CLA has write access to
	//
	// CLAscratch is the legacy code model scratch pad. use
	// .scratchpad in the new model
	CLAscratch		:
					{ *.obj(CLAscratch)
					. += CLA_SCRATCHPAD_SIZE;
					*.obj(CLAscratch_end) }
                     >  RAMLS4_5,  			PAGE = 1

	.scratchpad		: > RAMLS4_5,			PAGE = 1
	.bss_cla		: > RAMLS4_5,       	PAGE = 1
	.const_cla	    : > RAMLS4_5,			PAGE = 1


	//自定义字段
	ramgs0			: > RAMGS14,			PAGE = 1

	MSGRAM_CPU1_TO_CPU2	: >	CPU1TOCPU2RAM	PAGE = 1
	MSGRAM_CPU2_TO_CPU1	: >	CPU2TOCPU1RAM	PAGE = 1

	.reset			: > RESET,				PAGE = 0,	TYPE = DSECT //未用到
}

4.2 FLASH运行

CLA_SCRATCHPAD_SIZE = 0x100;

MEMORY
{
PAGE 0 :
	BEGIN           	: origin = 0x080000, length = 0x000002

	RAMGS0      		: origin = 0x00C000, length = 0x001000
	RAMGS1      		: origin = 0x00D000, length = 0x001000
	RAMGS2      		: origin = 0x00E000, length = 0x001000

	FLASHA				: origin = 0x080002, length = 0x001FFE	/* on-chip Flash */
	FLASHB				: origin = 0x082000, length = 0x002000	/* on-chip Flash */
	FLASHC				: origin = 0x084000, length = 0x002000	/* on-chip Flash */
	FLASHD				: origin = 0x086000, length = 0x002000	/* on-chip Flash */
	FLASHE				: origin = 0x088000, length = 0x008000	/* on-chip Flash */
	FLASHF				: origin = 0x090000, length = 0x008000	/* on-chip Flash */
	FLASHG				: origin = 0x098000, length = 0x008000	/* on-chip Flash */
	FLASHH				: origin = 0x0A0000, length = 0x008000	/* on-chip Flash */
	FLASHI				: origin = 0x0A8000, length = 0x008000	/* on-chip Flash */

	RESET           	: origin = 0x3FFFC0, length = 0x000002


PAGE 1 : /* Data Memory */
         /* Memory (RAM/FLASH) blocks can be moved to PAGE0 for program allocation */

	BOOT_RSVD         	: origin = 0x000002, length = 0x00004E   /* Part of M0, BOOT rom will use this for stack */

	RAMM0             	: origin = 0x000050, length = 0x0003B0
	RAMM1             	: origin = 0x000400, length = 0x000400

	RAMD0            	: origin = 0x00B000, length = 0x000800
	RAMD1            	: origin = 0x00B800, length = 0x000800

	RAMGS3      		: origin = 0x00F000, length = 0x001000
	RAMGS4      		: origin = 0x010000, length = 0x001000
	RAMGS5      		: origin = 0x011000, length = 0x001000
	RAMGS6      		: origin = 0x012000, length = 0x001000
	RAMGS7      		: origin = 0x013000, length = 0x001000
	RAMGS8				: origin = 0x014000, length = 0x001000
	RAMGS9				: origin = 0x015000, length = 0x001000
	RAMGS10				: origin = 0x016000, length = 0x001000
	RAMGS11				: origin = 0x017000, length = 0x001000
	RAMGS12				: origin = 0x018000, length = 0x001000
	RAMGS13				: origin = 0x019000, length = 0x001000
	RAMGS14				: origin = 0x01A000, length = 0x001000
	RAMGS15				: origin = 0x01B000, length = 0x001000

	RAMLS0_1			: origin = 0x008000, length = 0x002000
	RAMLS2_3			: origin = 0x00A000, length = 0x000A00
	RAMLS4_5			: origin = 0x00AA00, length = 0x000600

	FLASHJ				: origin = 0x0B0000, length = 0x008000	/* on-chip Flash */
	FLASHK				: origin = 0x0B8000, length = 0x002000	/* on-chip Flash */
	FLASHL				: origin = 0x0BA000, length = 0x002000	/* on-chip Flash */
	FLASHM				: origin = 0x0BC000, length = 0x002000	/* on-chip Flash */
	FLASHN				: origin = 0x0BE000, length = 0x001FF0	/* on-chip Flash */

	CPU2TOCPU1RAM		: origin = 0x03F800, length = 0x000400
	CPU1TOCPU2RAM		: origin = 0x03FC00, length = 0x000400

	CLA1_MSGRAMLOW		: origin = 0x001480, length = 0x000080
	CLA1_MSGRAMHIGH		: origin = 0x001500, length = 0x000080
}

SECTIONS
{
	codestart			: >	BEGIN									PAGE = 0,	ALIGN(8)

	.text				: >>FLASHB | FLASHC | FLASHD | FLASHE		PAGE = 0,	ALIGN(8)
	.cinit				: >	FLASHG									PAGE = 0,	ALIGN(8)
	.const				: > FLASHG									PAGE = 0,	ALIGN(8)
	.switch				: > FLASHH									PAGE = 0,	ALIGN(8)
	.init_array			: > FLASHH									PAGE = 0,	ALIGN(8)

	.stack				: > RAMM1									PAGE = 1
	.bss				: > RAMGS14,								PAGE = 1
	.bss:output			: > RAMGS14,								PAGE = 1
	.bss:cio			: > RAMGS14,								PAGE = 1
	.data				: > RAMGS14,								PAGE = 1
	.sysmem				: > RAMGS14,								PAGE = 1

	.TI.ramfunc : {}	LOAD = FLASHJ,
						RUN = RAMGS15,
						LOAD_START(RamfuncsLoadStart),
						LOAD_SIZE(RamfuncsLoadSize),
						LOAD_END(RamfuncsLoadEnd),
						RUN_START(RamfuncsRunStart),
						RUN_SIZE(RamfuncsRunSize),
						RUN_END(RamfuncsRunEnd),
						PAGE = 1, ALIGN(8)

   Cla1Prog:cla_dsp_fft : > RAMLS0_1, HIGH			PAGE = 1

	CLA1fftTables	:	LOAD = FLASHB,
						RUN = RAMLS4_5,
						RUN_START(CLA1fftTablesRunStart),
						LOAD_START(CLA1fftTablesLoadStart),
						LOAD_SIZE(CLA1fftTablesLoadSize),
						PAGE = 1


	Cla1Prog		:	LOAD = FLASHK,
						RUN = RAMLS0_1,
						RUN_START(Cla1ProgRunStart),
						LOAD_START(Cla1ProgLoadStart),
						LOAD_SIZE(Cla1ProgLoadSize),
						PAGE = 1

	CLAscratch		:
					{ *.obj(CLAscratch)
					. += CLA_SCRATCHPAD_SIZE;
					*.obj(CLAscratch_end) } >  RAMLS4_5,	PAGE = 1

	Cla1ToCpuMsgRAM			: > CLA1_MSGRAMLOW,				PAGE = 1
	CpuToCla1MsgRAM			: > CLA1_MSGRAMHIGH,			PAGE = 1

	.scratchpad				: > RAMLS4_5,					PAGE = 1
	.bss_cla				: > RAMLS4_5,       			PAGE = 1
	.const_cla				: > RAMLS4_5,       			PAGE = 1
	IOBuffer				: > RAMLS2_3,					PAGE = 1

	ramgs0					: > RAMGS13,					PAGE = 1

	Filter_RegsFile			: > RAMGS15,					PAGE = 1
	Filter1_RegsFile		: > RAMGS15,					PAGE = 1, fill=0x1111
	Filter2_RegsFile		: > RAMGS15,					PAGE = 1, fill=0x2222
	Filter3_RegsFile		: > RAMGS15,					PAGE = 1, fill=0x3333
	Filter4_RegsFile		: > RAMGS15,					PAGE = 1, fill=0x4444
	Difference_RegsFile		: > RAMGS15,					PAGE = 1, fill=0x3333

	MSGRAM_CPU1_TO_CPU2	: >	CPU1TOCPU2RAM					PAGE = 1
	MSGRAM_CPU2_TO_CPU1	: >	CPU2TOCPU1RAM					PAGE = 1

   GROUP : > CPU1TOCPU2RAM, PAGE = 1
    {
        PUTBUFFER
        PUTWRITEIDX
        GETREADIDX
    }

    GROUP : > CPU2TOCPU1RAM, PAGE = 1
    {
        GETBUFFER :    TYPE = DSECT
        GETWRITEIDX :  TYPE = DSECT
        PUTREADIDX :   TYPE = DSECT
    }

    .reset				: > RESET,							PAGE = 0,	TYPE = DSECT /* not used, */
}

上述程序我就不解释了，如果需要我可以再更新，大家可以留言

5. 运行程序

初始化后，我们如何使用CLA呢？在TI的管理中有一个文件后缀为 xxx.CLA，我们在里面写CLA的运行程序，但是它不同于其他C文件，因为它不允许定义变量，不允许被include，只能写函数和include其他文件。

//
// Included Files
//
#include "CLA_share.h"
#include "cla_cfft.h"
#include "CLAmath.h"

extern float fVal;           //Holds the input argument to the task
extern float fResult;        //The arsine of the input argument

__interrupt void Cla1Task1 ( void )
{
    __mdebugstop();
    CLA_CFFT_run256Pt();
    fResult=
                CLAsin(4.2);
}

interrupt void Cla1Task2 ( void )
{
    fResult=
                    CLAsin(0.5)*4096;
}

interrupt void Cla1Task3 ( void )
{

}

interrupt void Cla1Task4 ( void )
{

}

interrupt void Cla1Task5 ( void )
{

}

interrupt void Cla1Task6 ( void )
{

}

interrupt void Cla1Task7 ( void )
{

}

interrupt void Cla1Task8 ( void )
{

}

这里我就用Task1来介绍了，大家可以先将CLA_CFFT_run256Pt();注释，这是256点的FFT程序。我们先使用CLAsin来做。其作用与math中的sin一致给一个数程序算出sin值，它是用弧度值来计算。

除了上述的CLA文件还有配套的C文件非常重要。所有的中断，

/*
 * CLA_interrupt.c
 *
 *  Created on: 2023年5月1日
 *      Author: 18752
 */

#include "CLA_interrupt.h"

//
// cla1Isr1 - CLA1 ISR 1
//

__interrupt void cla1Isr1 ()
{
    //
    // Acknowledge the end-of-task interrupt for task 1
    //
    Interrupt_clearACKGroup(INTERRUPT_ACK_GROUP11);

    //
    // Uncomment to halt debugger and stop here
    //
    // asm(" ESTOP0");
}

//
// cla1Isr2 - CLA1 ISR 2
//
__interrupt void cla1Isr2 ()
{
    asm(" ESTOP0");
}

//
// cla1Isr3 - CLA1 ISR 3
//
__interrupt void cla1Isr3 ()
{
    asm(" ESTOP0");
}

//
// cla1Isr4 - CLA1 ISR 4
//
__interrupt void cla1Isr4 ()
{
    asm(" ESTOP0");
}

//
// cla1Isr5 - CLA1 ISR 5
//
__interrupt void cla1Isr5 ()
{
    asm(" ESTOP0");
}

//
// cla1Isr6 - CLA1 ISR 6
//
__interrupt void cla1Isr6 ()
{
    asm(" ESTOP0");
}

//
// cla1Isr7 - CLA1 ISR 7
//
__interrupt void cla1Isr7 ()
{
    asm(" ESTOP0");
}

//
// cla1Isr8 - CLA1 ISR 8
//
__interrupt void cla1Isr8 ()
{
    //
    // Acknowledge the end-of-task interrupt for task 8
    //
    Interrupt_clearACKGroup(INTERRUPT_ACK_GROUP11);
    //
    // Uncomment to halt debugger and stop here
    //
//    asm(" ESTOP0");
}

这是CLA的中断文件，我就过多介绍了，抄就完了。

在上述内容之外你还需要找一个C文件定义你的变量，如果想让cla使用则必须先对区域定义，相信大家看到就知道是怎么用的了。

#pragma DATA_SECTION(fVal,"CpuToCla1MsgRAM");
float fVal;
#pragma DATA_SECTION(fResult,"Cla1ToCpuMsgRAM");
float fResult;

CPUtoCLA即CPU的变量放到CLA中，另一个就是CLAtoCPU了。

5.库文件

对于CLA还要依赖的lib文件，其路径是两个。

C2000Ware_4_03_00_00\libraries\boot_rom\f2837xd\rev0\rom_symbol_libs

C2000Ware_4_03_00_00\libraries\math\CLAmath\c28\lib

6606f5c76adf4c33a7156dc85493b2b1.png

 如果你使用的是elf文件就和我一样配置，如果是legacy可以看文件的后缀开启你用的文件

6.最后移植完成后的文件

下图是文件结构

0d2ffab828ab4f7ca2163947c85b6b2d.png

CLA.init

/*
 * CLA_init.c
 *
 *  Created on: 2023年5月1日
 *      Author: 18752
 */

#include "CLA_init.h"



#pragma DATA_SECTION(fVal,"CpuToCla1MsgRAM");
float fVal;
#pragma DATA_SECTION(fResult,"Cla1ToCpuMsgRAM");
float fResult;

#pragma DATA_SECTION(IOBuffer,"IOBuffer")
float IOBuffer[(256+1)*2] = {
    #include "ffttest.h"
};


//
// CLA_runTest - Execute CLA task tests for specified vectors
//
void CLA_runTest(unsigned char RTest)
{
    switch(RTest)
    {
        case 1: CLA_forceTasks(CLA1_BASE,CLA_TASKFLAG_1);   break;
        case 2: CLA_forceTasks(CLA1_BASE,CLA_TASKFLAG_2);   break;
    }
    WAITSTEP;
}

void CLA_init()
{
    EALLOW;

    CLA_DSP_initEpie();

    CLA_DSP_configClaMemory();

    CLA_DSP_initCpu1Cla1();

    EDIS;
}

void CLA_DSP_initEpie()
{
    Interrupt_initModule();
    Interrupt_initVectorTable();
}


void CLA_DSP_configClaMemory(void)
{
    extern uint32_t Cla1ProgRunStart, Cla1ProgLoadStart, Cla1ProgLoadSize;
    extern uint32_t CLA1fftTablesRunStart, CLA1fftTablesLoadStart, CLA1fftTablesLoadSize;

    EALLOW;

#ifdef _FLASH
    // Copy over code and tables from FLASH to RAM
    memcpy((uint32_t *)&Cla1ProgRunStart, (uint32_t *)&Cla1ProgLoadStart,
            (uint32_t)&Cla1ProgLoadSize);

    memcpy((uint32_t *)&CLA1fftTablesRunStart, (uint32_t *)&CLA1fftTablesLoadStart,
            (uint32_t)&CLA1fftTablesLoadSize);
#endif //_FLASH

    // Initialize and wait for CLA1ToCPUMsgRAM
    MemCfg_initSections(MEMCFG_SECT_MSGCLA1TOCPU);
    while (!MemCfg_getInitStatus(MEMCFG_SECT_MSGCLA1TOCPU)){};

    // Initialize and wait for CPUToCLA1MsgRAM
    MemCfg_initSections(MEMCFG_SECT_MSGCPUTOCLA1);
    while (!MemCfg_getInitStatus(MEMCFG_SECT_MSGCPUTOCLA1)){};

    // Select LS0-LS1RAM to be the programming space for the CLA
    // First configure the CLA to be the master for LS0 and then
    // set the space to be a program block
    MemCfg_setLSRAMMasterSel(MEMCFG_SECT_LS0,MEMCFG_LSRAMMASTER_CPU_CLA1);
    MemCfg_setCLAMemType(MEMCFG_SECT_LS0,MEMCFG_CLA_MEM_PROGRAM);
    MemCfg_setLSRAMMasterSel(MEMCFG_SECT_LS1,MEMCFG_LSRAMMASTER_CPU_CLA1);
    MemCfg_setCLAMemType(MEMCFG_SECT_LS1,MEMCFG_CLA_MEM_PROGRAM);

    //Next configure LS2-LS5RAM as data spaces for the CLA
    // First configure the CLA to be the master for LSx and then
    // set the spaces to be code blocks

    MemCfg_setLSRAMMasterSel(MEMCFG_SECT_LS2,MEMCFG_LSRAMMASTER_CPU_CLA1);
    MemCfg_setCLAMemType(MEMCFG_SECT_LS2, MEMCFG_CLA_MEM_DATA);

    MemCfg_setLSRAMMasterSel(MEMCFG_SECT_LS3,MEMCFG_LSRAMMASTER_CPU_CLA1);
    MemCfg_setCLAMemType(MEMCFG_SECT_LS3, MEMCFG_CLA_MEM_DATA);

    MemCfg_setLSRAMMasterSel(MEMCFG_SECT_LS4,MEMCFG_LSRAMMASTER_CPU_CLA1);
    MemCfg_setCLAMemType(MEMCFG_SECT_LS4, MEMCFG_CLA_MEM_DATA);

    MemCfg_setLSRAMMasterSel(MEMCFG_SECT_LS5,MEMCFG_LSRAMMASTER_CPU_CLA1);
    MemCfg_setCLAMemType(MEMCFG_SECT_LS5, MEMCFG_CLA_MEM_DATA);

}

void CLA_DSP_initCpu1Cla1(void)
{
    // Compute all CLA task vectors
    // On Type-1 CLAs the MVECT registers accept full 16-bit task addresses as
    // opposed to offsets used on older Type-0 CLAs

//
// Suppressing #770-D conversion from pointer to smaller integer
// The CLA address range is 16 bits so the addresses passed to the MVECT
// registers will be in the lower 64KW address space. Turn the warning
// back on after the MVECTs are assigned addresses
//
#pragma diag_suppress=770
    CLA_mapTaskVector(CLA1_BASE,CLA_MVECT_1,(uint16_t)&Cla1Task1);
    CLA_mapTaskVector(CLA1_BASE,CLA_MVECT_2,(uint16_t)&Cla1Task2);
    CLA_mapTaskVector(CLA1_BASE,CLA_MVECT_3,(uint16_t)&Cla1Task3);
    CLA_mapTaskVector(CLA1_BASE,CLA_MVECT_4,(uint16_t)&Cla1Task4);
    CLA_mapTaskVector(CLA1_BASE,CLA_MVECT_5,(uint16_t)&Cla1Task5);
    CLA_mapTaskVector(CLA1_BASE,CLA_MVECT_6,(uint16_t)&Cla1Task6);
    CLA_mapTaskVector(CLA1_BASE,CLA_MVECT_7,(uint16_t)&Cla1Task7);
    CLA_mapTaskVector(CLA1_BASE,CLA_MVECT_8,(uint16_t)&Cla1Task8);

#pragma diag_warning=770

    // Enable IACK instruction to start a task on CLA in software
    // for all  8 CLA tasks

    CLA_enableIACK(CLA1_BASE);
    CLA_enableTasks(CLA1_BASE, CLA_TASKFLAG_ALL);

    // Configure the vectors for the end-of-task interrupt for all
    // 8 tasks

    Interrupt_register(INT_CLA1_1, &cla1Isr1);
    Interrupt_register(INT_CLA1_2, &cla1Isr2);
    Interrupt_register(INT_CLA1_3, &cla1Isr3);
    Interrupt_register(INT_CLA1_4, &cla1Isr4);
    Interrupt_register(INT_CLA1_5, &cla1Isr5);
    Interrupt_register(INT_CLA1_6, &cla1Isr6);
    Interrupt_register(INT_CLA1_7, &cla1Isr7);
    Interrupt_register(INT_CLA1_8, &cla1Isr8);
}

CLA_init.h

/*
 * CLA_init.h
 *
 *  Created on: 2023年5月1日
 *      Author: 18752
 */

#ifndef CLA_CLA_INIT_H_
#define CLA_CLA_INIT_H_


//
// Included Files
//

#include "driverlib.h"
#include "device.h"


#define FFTLENGTH 256


//*****************************************************************************
//
// CLA Configurations
//
//*****************************************************************************
#define myCLA0_BASE CLA1_BASE

//
// The following are symbols defined in the CLA assembly code
// Including them in the shared header file makes them global
// and the main CPU can make use of them.
//
__attribute__((interrupt)) void Cla1Task1();
__attribute__((interrupt)) void Cla1Task2();
__attribute__((interrupt)) void Cla1Task3();
__attribute__((interrupt)) void Cla1Task4();
__attribute__((interrupt)) void Cla1Task5();
__attribute__((interrupt)) void Cla1Task6();
__attribute__((interrupt)) void Cla1Task7();
__attribute__((interrupt)) void Cla1Task8();
void myCLA0_init();


//*****************************************************************************
//
// INTERRUPT Configurations
//
//*****************************************************************************

// Interrupt Settings for INT_myCLA01
#define INT_myCLA01 INT_CLA1_1
#define INT_myCLA01_INTERRUPT_ACK_GROUP INTERRUPT_ACK_GROUP11
extern __interrupt void cla1Isr1(void);

// Interrupt Settings for INT_myCLA02
#define INT_myCLA02 INT_CLA1_2
#define INT_myCLA02_INTERRUPT_ACK_GROUP INTERRUPT_ACK_GROUP11
extern __interrupt void cla1Isr2(void);

// Interrupt Settings for INT_myCLA03
#define INT_myCLA03 INT_CLA1_3
#define INT_myCLA03_INTERRUPT_ACK_GROUP INTERRUPT_ACK_GROUP11
extern __interrupt void cla1Isr3(void);

// Interrupt Settings for INT_myCLA04
#define INT_myCLA04 INT_CLA1_4
#define INT_myCLA04_INTERRUPT_ACK_GROUP INTERRUPT_ACK_GROUP11
extern __interrupt void cla1Isr4(void);

// Interrupt Settings for INT_myCLA05
#define INT_myCLA05 INT_CLA1_5
#define INT_myCLA05_INTERRUPT_ACK_GROUP INTERRUPT_ACK_GROUP11
extern __interrupt void cla1Isr5(void);

// Interrupt Settings for INT_myCLA06
#define INT_myCLA06 INT_CLA1_6
#define INT_myCLA06_INTERRUPT_ACK_GROUP INTERRUPT_ACK_GROUP11
extern __interrupt void cla1Isr6(void);

// Interrupt Settings for INT_myCLA07
#define INT_myCLA07 INT_CLA1_7
#define INT_myCLA07_INTERRUPT_ACK_GROUP INTERRUPT_ACK_GROUP11
extern __interrupt void cla1Isr7(void);

// Interrupt Settings for INT_myCLA08
#define INT_myCLA08 INT_CLA1_8
#define INT_myCLA08_INTERRUPT_ACK_GROUP INTERRUPT_ACK_GROUP11
extern __interrupt void cla1Isr8(void);

//*****************************************************************************
//
// MEMCFG Configurations
//
//*****************************************************************************

void CLA_init();
// \brief Initialize system clocks
//
void CLA_DSP_initSystemClocks(void);

// \brief Initialize Enhanced PIE
//
void CLA_DSP_initEpie(void);

// \brief Configure CLA memory space
//
void CLA_DSP_configClaMemory(void);

// \brief Intialize the MVECT registers for the CLA
//
void CLA_DSP_initCpu1Cla1(void);
void CLA_runTest(unsigned char RTest);

#define WAITSTEP     asm(" RPT #255 || NOP")

#endif /* CLA_CLA_INIT_H_ */

CLA_initerrupt.c

/*
 * CLA_interrupt.c
 *
 *  Created on: 2023年5月1日
 *      Author: 18752
 */

#include "CLA_interrupt.h"

//
// cla1Isr1 - CLA1 ISR 1
//

__interrupt void cla1Isr1 ()
{
    //
    // Acknowledge the end-of-task interrupt for task 1
    //
    Interrupt_clearACKGroup(INTERRUPT_ACK_GROUP11);

    //
    // Uncomment to halt debugger and stop here
    //
    // asm(" ESTOP0");
}

//
// cla1Isr2 - CLA1 ISR 2
//
__interrupt void cla1Isr2 ()
{
    asm(" ESTOP0");
}

//
// cla1Isr3 - CLA1 ISR 3
//
__interrupt void cla1Isr3 ()
{
    asm(" ESTOP0");
}

//
// cla1Isr4 - CLA1 ISR 4
//
__interrupt void cla1Isr4 ()
{
    asm(" ESTOP0");
}

//
// cla1Isr5 - CLA1 ISR 5
//
__interrupt void cla1Isr5 ()
{
    asm(" ESTOP0");
}

//
// cla1Isr6 - CLA1 ISR 6
//
__interrupt void cla1Isr6 ()
{
    asm(" ESTOP0");
}

//
// cla1Isr7 - CLA1 ISR 7
//
__interrupt void cla1Isr7 ()
{
    asm(" ESTOP0");
}

//
// cla1Isr8 - CLA1 ISR 8
//
__interrupt void cla1Isr8 ()
{
    //
    // Acknowledge the end-of-task interrupt for task 8
    //
    Interrupt_clearACKGroup(INTERRUPT_ACK_GROUP11);
    //
    // Uncomment to halt debugger and stop here
    //
//    asm(" ESTOP0");
}

CLA_initerrupt.h

/*
 * CLA_interrupt.h
 *
 *  Created on: 2023年5月1日
 *      Author: 18752
 */

#ifndef CLA_CLA_INTERRUPT_H_
#define CLA_CLA_INTERRUPT_H_


#include "interrupt.h"


#endif /* CLA_CLA_INTERRUPT_H_ */

CLA_share.h

/*
 * CLA_share.h
 *
 *  Created on: 2023年5月1日
 *      Author: 18752
 */

#ifndef CLA_CLA_SHARE_H_
#define CLA_CLA_SHARE_H_

#include <stdint.h>


//
// Defines
//
#define PI                3.141592653589


extern float fVal;
extern float fResult;


__interrupt void Cla1Task1();
__interrupt void Cla1Task2();
__interrupt void Cla1Task3();
__interrupt void Cla1Task4();
__interrupt void Cla1Task5();
__interrupt void Cla1Task6();
__interrupt void Cla1Task7();
__interrupt void Cla1Task8();





#endif /* CLA_CLA_SHARE_H_ */

MyCLA.cla

//
// Included Files
//
#include "CLA_share.h"
#include "cla_cfft.h"
#include "CLAmath.h"

extern float fVal;           //Holds the input argument to the task
extern float fResult;        //The arsine of the input argument

__interrupt void Cla1Task1 ( void )
{
    __mdebugstop();
    CLA_CFFT_run256Pt();
    fResult=
                CLAsin(4.2);
}

interrupt void Cla1Task2 ( void )
{
    fResult=
                    CLAsin(0.5)*4096;
}

interrupt void Cla1Task3 ( void )
{

}

interrupt void Cla1Task4 ( void )
{

}

interrupt void Cla1Task5 ( void )
{

}

interrupt void Cla1Task6 ( void )
{

}

interrupt void Cla1Task7 ( void )
{

}

interrupt void Cla1Task8 ( void )
{

}

7. 最后效果

46d2e0cf22fc4d5b8a5aeb9f2bbc6118.jpg

0362af62d0084f26afa0477baa88cd76.png

可以看到cla初始化后fresult是0

afc343e880bd4072a42222df98ffbb47.png

fe5febd9d431443398c1bac08b808c2a.png

 cla运行完成后是0.707与计算器算的是一样的

# dsp开发
Logo
开放原子开发者工作坊

开放原子开发者工作坊旨在鼓励更多人参与开源活动，与志同道合的开发者们相互交流开发经验、分享开发心得、获取前沿技术趋势。工作坊有多种形式的开发者活动，如meetup、训练营等，主打技术交流，干货满满，真诚地邀请各位开发者共同参与！

更多推荐

“源”聚天大，码动未来|开放原子校源行（天津大学站）即将启幕！

11月22日，开放原子校源行（天津大学站）将在天津大学北洋园校区隆重举办。

avatar 开放原子开发者工作坊

openEuler系累计装机量突破1000万，树立操作系统产业新里程碑

11月15日，以“以智能，致世界”为主题的操作系统大会2024在北京中关村国际创新中心召开。

avatar 开放原子开发者工作坊

多模态大模型&科学计算双管齐下，百度飞桨两大赛项报名倒计时！

第二届开放原子大赛是由开放原子开源基金会组织举办的开源技术领域专业赛事，聚焦解决真问题，重点覆盖基础软件、工业软件、人工智能大模型等领域

avatar 开放原子开发者工作坊