TMS320F28379D: Memory Test Question

Any updates on this?

Thanks,
Chuchen

Hi Chuchen,

How were you observing an ECC error? Did you do CPU reads to those location (406 and 408)?

Regards,
Nirav

HI Nirav,

Thanks for the response. I've asked the customer to provide feedback to your questions here.

Thanks,
Chuchen

Hi Chuchen,

It would be good if we can get the test code they are using to see the full procedure they are following for this test.

Regards,
Vivek Singh

Hi .. The following is my test code.  As noted in the last section, I am using the debugger to see which registers indicate memory test error: i was monitoring MemErrorRegs as well as NMI registers.  I saw no change.

The code:

// Make sure that RAM is OK
#define numberOfLocsToTest 5
volatile uint16* ptr = (uint16*)0x404;

for (int i = 0; i < numberOfLocsToTest; i++)
{
// initialize test area to zero.
ptr[i] = 0;
}
// Enable test register.
EALLOW;
MemCfgRegs.DxTEST.s.TEST_M1 = 0x01; // Writes are allowed to data bits only. No write to ECC bits
EDIS;
// Write failures.. combination of 1 and 2 bits.
for (int i = 0; i < numberOfLocsToTest; i++)
{
// initialize test area to zero.
ptr[i] = i+1;
}
// Read back data
uint16 readBack[numberOfLocsToTest];

for (int i = 0; i < numberOfLocsToTest; i++)
{
// should cause error bits to set..
// I expect NMI and other TBD error registers to trigger..
// As a test... use the debugger to figure out which registers trigger.

// NOTE: NOTHING HAPPENS!!
readBack[i] = ptr[i];
}

EALLOW;
MemCfgRegs.DxTEST.s.TEST_M1 = 0x00; // Functional mode.
EDIS;

Hi,

You need to exit the TEST MODE and then read the value to generate error. Basically after you are done with data corruption, change the TEST_M1 bit value to 0x00 and then do the read operation. Please try the same and let us know how it goes.

Regards,

Vivek Singh