TMS570LS3137: Testing the interrupt priority remapping

Hello all,

I wanted to test the remapping of the interrupt by using the function vim. To evaluate the effectiveness of the remapping I thought about the following scenario:

1) The RTI-interrupt drives periodically the status of a GIO output to be True for 100 ms and False for the next 100 ms. The result is a square wave with 200ms of period.

2) A big computational load is executed when a GIO input goes to True. The computation consists in a 10M for-loop empty iterations.

3) Initially a higher priority is given to the GIO input (and then to the associated for loop) than the RTI-interrupt.

The effect of the above described scenario is having a 200ms square wave that gets stacked to the last logical value for approx 1 second when the GIO input occurs. It works and the code is the following:

#include "sys_common.h"

/* USER CODE BEGIN (1) */
#include "system.h"
#include "sys_vim.h"
#include "het.h"
#include "gio.h"
#include "rti.h"
/* USER CODE END */



/* USER CODE BEGIN (2) */
uint32 x;
/* USER CODE END */

int main(void)
{
/* USER CODE BEGIN (3) */

    vimInit();
    _enable_IRQ();
    vimEnableInterrupt(2,SYS_IRQ);
    vimEnableInterrupt(23,SYS_IRQ);

/*Swap the default priority of the RTI with the default priority of the GIO(B) input */
    vimChannelMap(2,23,&rtiCompare0Interrupt);
    vimChannelMap(23,2,&gioLowLevelInterrupt);

    rtiInit();
    rtiEnableNotification(rtiNOTIFICATION_COMPARE0);
    gioInit();
    gioEnableNotification(gioPORTB, 2);
    gioSetBit(gioPORTB,3,1);
    rtiStartCounter(rtiCOUNTER_BLOCK0);
/* USER CODE END */

    return 0;
}
/* USER CODE BEGIN (4) */

/*To generate the square wave*/

void rtiNotification(uint32 notification)
{
    x++;
    if (x % 2 == 1)
    {
        gioSetBit(gioPORTA,4,1);
    }
    else
    {
        gioSetBit(gioPORTA,4,0);
    }
}

/*computational load (approx 1s delay)*/
void gioNotification(gioPORT_t *port, uint32 bit)
{
    if (port == gioPORTB)
    {
        uint32 i = 0;
        for (i=0;i<10000000;i++);
    }
}

In Halcogen, aside from changing the period of the RTI compare 0 to 100ms and setting the GIO output intended to use, the settings in the screenshots were applied to the VIM tab and the GIO input used as "external" source of interrupt:

Now, I would like to resume the original priority with the RTI over the GIO, and expect that even if the input occurs the square wave shall "kill" it at the occurrence of the 200ms square wave, than the only part that I change in the code is the following:

vimChannelMap(2,2,&rtiCompare0Interrupt);
vimChannelMap(23,23,&gioLowLevelInterrupt);

but unfortunately it produces the same effect: again, the 200ms square wave is delayed by the for-loop at the occurrence of the input. What's wrong in the priority assignment or in the test scenario?

Thanks ahead for any reply.

Regards.