topic Re: Strange behaviour of Safe Routine for writting flash in Digital Signal Controllers

Strange behaviour of Safe Routine for writting flash

cabl — Tue, 21 Dec 2021 07:32:03 GMT

Hi everybody,

I´m using a 56F82646 part and having some trouble with the PE´s IntFlash component. To my understanding there´s something odd in the PE code for the SafeRoutineCaller:

- when stack location where SaveRoutineStackSpace is placed is under 0x7FF (RAM address), SafeRoutineInDataPtr gets a value under 0xFFF (double of SaveRoutineStackSpace address), and SafeRoutineInCodePtr=0xF000 + SafeRoutineInDataPtr is a valid address under 0xFFFF and everything works fine. At least this es what I can see in the debugger.

- when stack location where SaveRoutineStackSpace is placed is over 0x7FF, SafeRoutineInDataPtr gets a value over 0xFFF and SafeRoutineInCodePtr=0xF000 + SafeRoutineInDataPtr is a INVALID address (over 0xFFFF). Now it does not work.

If I force SaveRoutineStackSpace to be global and be placed at low RAM addresses it always work fine, without any other change in the code, just the SaveRoutineStackSpace position.

This behaviour seems to be independent of the memory model and width of the pointer used in the project options.

Am I not understanding something here?

Thanks in advance.

Re: Strange behaviour of Safe Routine for writting flash

xiangjun_rong — Thu, 23 Dec 2021 03:20:49 GMT

Hi,

As the following screenshot in MCU_DSC_Compiler.pdf, if you use byte access in the application code and in small model, the address must be less than 0x7FFF.

The MCU_DSC_Compiler.pdf is located

......\Freescale\CW MCU v11.1\MCU\Help\PDF

Hope it can help you

XiangJun Rong

Re: Strange behaviour of Safe Routine for writting flash

cabl — Thu, 23 Dec 2021 06:24:35 GMT

Thanks for the response! I did know that. But the problem is I can´t see any warning or protection in the PE IntFlash component code that protects against invalid address when there is high stack usage. How should this code be modified to force word access?

Thanks

Re: Strange behaviour of Safe Routine for writting flash

xiangjun_rong — Thu, 23 Dec 2021 09:20:29 GMT

Hi,

Pls try to use large data model rather than small data model to avoid the issue.

XiangJun Rong

Re: Strange behaviour of Safe Routine for writting flash

cabl — Mon, 27 Dec 2021 07:03:51 GMT

No luck, sorry. The large memory model still fails when executing SafeRoutineCaller if stack usage is high.

Re: Strange behaviour of Safe Routine for writting flash

xiangjun_rong — Tue, 28 Dec 2021 07:56:19 GMT

Hi,

I have reviewed your code, but I do not see the code to erase flash, if you program flash space which are not erased, you may damage the flash.

How about using the code:

/*!
** @file main.c
** @version 01.16
** @brief
** Main module.
** This module contains user's application code.
*/
/*!
** @addtogroup main_module main module documentation
** @{
*/
/* MODULE main */

/* Including needed modules to compile this module/procedure */
#include "Cpu.h"
#include "Events.h"
#include "Pins1.h"
#include "IFsh1.h"
#include "IntFlashLdd1.h"
/* Including shared modules, which are used for whole project */
#include "PE_Types.h"
#include "PE_Error.h"
#include "PE_Const.h"
#include "IO_Map.h"
#include "Init_Config.h"
#include "PDD_Includes.h"

#define PROADDRESS 0x7000

unsigned int array[1024];
unsigned int temp,i;
unsigned int *pointer;

void main(void)
{
/* Write your local variable definition here */

/*** Processor Expert internal initialization. DON'T REMOVE THIS CODE!!! ***/
PE_low_level_init();
/*** End of Processor Expert internal initialization. ***/

/* Write your code here */
/* Write your code here */
for(i=0;i<1024;i++)
{
array[i]=i;
}
pointer=(unsigned int *)PROADDRESS;
//erase program flash from address:0x1 F000
IFsh1_EraseSector((unsigned long)PROADDRESS);
pointer=(unsigned int *)PROADDRESS;
asm(nop);
for(i=0; i<1024; i++)
{
IFsh1_SetWordFlash((unsigned long)pointer,array[i]);
pointer++;
}
for(;;) {}
}

/* END main */
/*!
** @}
*/
/*
** ###################################################################
**
** This file was created by Processor Expert 10.3 [05.09]
** for the Freescale 56800 series of microcontrollers.
**
** ###################################################################
*/

XiangJun Rong

Re: Strange behaviour of Safe Routine for writting flash

cabl — Mon, 03 Jan 2022 09:43:57 GMT

Hi!

The erasure of the sector is done automatically by the PE component if needed, no problem with this (option safe write with erase).

Have you been able to test IFsh1_SetWordFlash function when stack usage is high?

Regards,

Re: Strange behaviour of Safe Routine for writting flash

cabl — Mon, 31 Jan 2022 08:12:29 GMT

No chance of testing the "bug" (or my missinterpretation) in IFsh1_SetWordFlash?