TMS570LS12x Launchpad SafeTI Integration

Hello,

Thank you very much for all the answers.

I have another question related to the change from HALCoGen code to SafeTiCode.

I have generated the HAL code using HALCoGen.

But if I change the sys_startup.c file's

_coreInitRegisters_();

function to 

SL_Init_R4Registers();

The ERR LED in Launchpad becomes ON.

Why is this happening? What are the limitations? Are these limitations documented somewhere. 

When I checked the HALCoGen examples. It is also not changing _coreInitRegisters_() function. 

But it does not explain any reason.

I get 0x08 in the ESMSR3. I cannot find that bit definition in the data sheet. What is the error reported there?

In the sys_startup.c code, do we need to call _coreInitRegisters_()? Can't we just replace it with SL_Init_R4Registers()?

Pradeepa,

0x08 in ESMSR3 is documented:

I bet you were just looking at bit 8 instead, because bit 8 is reserved.   The value 0x08 means bit 3 is set.  Easy mistake to make.

Regarding your 2nd question - Yes I think you can but I would say that you need to make sure that you have defined the default initialization value that

SL_Init_R4Registers() uses.   I believe the other init routine just uses 0.

Now in my personal opinion - and I am filing a ticket on this - there is an issue in the startup code in that it tests the reset source before calling the core init registers function.   It may be 'ok' but it's pretty fragile as calling functions before the stack pointers are initialized isn't a good practice IMO.    It may work on a particular compiler version / set of switches but anyway... 

Point is if I were you for now I would put a call to _coreInitRegisters_ in the /* User Code */ right before the switch/case section that checks the reset state.
It may initialize the registers twice this way but it shouldn't hurt anything.

Anthony,

Yes I have missed the definition of the error.

I believe SL_Init_R4Registers() function uses the assembly code available in "sl_asm_api.asm". That assembly code uses SL_REG_INIT_VAL #define internally to initialize all the registers to zero. SL_REG_INIT_VAL is defined in sl_config.h.

All the above are verified in my code.

Also, following definitions are added to the project.

_VFP_SUPPORT_=1

_TMS570LS12x_

EXTERNAL_SP_INIT

Now I am using the example code available in "C:\ti\Hercules\HALCoGen\v04.05.01\examples\TMS570LS12x_11x\example_SafetyLib.c"

Example code as it is works perfectly. But I want to have a better understanding of the behavior of the library before using it in the project. Otherwise I am responsible for any failures.

I did very minor change to the code in _c_int00() function declared in sys_startup.c file.

The change is,

void _c_int00(void)
{
	/* NOTE: DO NOT INITIALIZE ANY VARIABLES !! (will cause DABORT because of un-initialized SP) */
	volatile boolean 			retVal;             /* For function return values */
    SL_STC_Config 				stcSelfTestConfig;  /* STC Configuration */

    _coreInitRegisters_();

    SL_Init_R4Registers();   // !!!! ****THIS IS THE ONLY CHANGE**** !!!!

    _coreInitStackPointer_();

Now I connected my TMS570LS12x launchpad and used the CCS debug feature to download the code to the internal flash.

Now I can see the ERROR RED LED is on. I paused the debugger and checked the ESM registers. I can see 0x08 in ESMSR3.

Then I changed the code like below,

void _c_int00(void)
{
	/* NOTE: DO NOT INITIALIZE ANY VARIABLES !! (will cause DABORT because of un-initialized SP) */
	volatile boolean 			retVal;             /* For function return values */
    SL_STC_Config 				stcSelfTestConfig;  /* STC Configuration */

    SL_Init_R4Registers();

    _coreInitRegisters_();
    
    _coreInitStackPointer_();

Now it is working. So it is evident that some issue is there in SL_Init_R4Registers function. 

Please check into this.

Thanks for the help so far.

PS: I tried to be as descriptive as possible to allow you to debug the issue quickly.

Hello Vivs,

Vivs Shitole said:

 I guess calling SL_Init_R4Registers() after _coreInitRegisters_() is breaking the system for some reason. It will work fine if you call SL_Init_R4Registers() only.

I can confirm that calling the function SL_Init_R4Registers() alone will still break the system. 

I have tried this several times. I will check this again today and let you know.

Thanks.

Case 1:

SL_Init_R4Registers();
/* USER CODE END */

/* Initialize Core Registers to avoid CCM Error */
_coreInitRegisters_();

/* USER CODE BEGIN (6) */
/* USER CODE END */

/* Initialize Stack Pointers */
_coreInitStackPointer_();

/* USER CODE BEGIN (7) */

No issue.

Case 2:

/* USER CODE END */

/* Initialize Core Registers to avoid CCM Error */
_coreInitRegisters_();

/* USER CODE BEGIN (6) */
SL_Init_R4Registers();
/* USER CODE END */

/* Initialize Stack Pointers */
_coreInitStackPointer_();

/* USER CODE BEGIN (7) */

Error LED is on

Case 3:

/* USER CODE END */

/* Initialize Core Registers to avoid CCM Error */
//_coreInitRegisters_();

/* USER CODE BEGIN (6) */
SL_Init_R4Registers();
/* USER CODE END */

/* Initialize Stack Pointers */
_coreInitStackPointer_();

/* USER CODE BEGIN (7) */

Error LED is on

Case 4:

/* USER CODE END */

/* Initialize Core Registers to avoid CCM Error */
_coreInitRegisters_();

/* USER CODE BEGIN (6) */
/* USER CODE END */

/* Initialize Stack Pointers */
_coreInitStackPointer_();

/* USER CODE BEGIN (7) */
SL_Init_R4Registers();

No Error LED. But the execution is going through a boot loop.

Hope this will help you.

Hello,

When will the next release be available?

Is there any document which contains the current bugs and workarounds?