TMS570LC4357: Memory test fails when run in standalone mode

Hi Gobind Singh,

Gobind Singh1 said:

But as soon as it is power cycled and allowed to run standalone, the memory test fails and it gets stuck in "testFail" assembly instruction. (Screenshot below)

How do you know where it stuck without JTAG?

--
Thanks & regards,
Jagadish.

I know it because:

1. After power cycle, I launched the ccxml file, selected "connect target" and loaded symbols (without reprogramming). It shows me where the code is stuck. Its the screenshot that I have shared in my first post. 

2. After step 1, if I do a soft reset and try to rerun, the test fails. It fails to write the pattern in the memory. Some of the memory is not even writable in the "Memory" window of the CCS.

3. When I reprogram the code and run it (without power cycle), it runs fine.

Hi Jagadish, 

Any update on this issue please?

Thank you.

Gobind Singh

Hi Gobind Singh,

Apologies for the delay, i was stuck with other issues. I will try to provide you an update in this week.

--
Thanks & regards,
Jagadish.

Hi Gobind Singh,

My sincere apologies for the delay in late response.

Are you still stuck with this issue?

--
Thanks & regards,
Jagadish.

Hi Jagadish,

Yes, I still have this problem.

Thanks

Gobind Singh

Hi Gobind Singh,

Sincere apologies for further delay.

Today i got a chance to debug your project. 

But here i could not face the issue you are talking about.

This is the execution result in debugger:

As we can see the code is stuck at the while(1) in main loop and which is assembly instruction at address 0x46AC.

What i did is now, i exit from the debugging mode and did power on reset of the board and after that i connected to the board without re-flashing the code to verify where code stuck.

Here is the result:

It is executing 0x46AC instruction only that means the code again stuck at the while(1) in main loop.

I am not facing the issue you are talking about.

--
Thanks & regards,
Jagadish.

Hi Jagadish,

Did you disconnect the USB power and reconnect or just used the switch on the board to do a reset? I think it works OK when we use the switch but not when we do a power cycle

Thank you

Gobind Singh

Hi Gobind Singh,

You are correct, i could see the issue after disconnecting and reconnecting USB cable:

Even if i try to write directly to the memory will also creating issue. I am suspecting maybe memory permissions issue causing this problem. As this code is not generated from our HALCoGen i couldn't be able to find root cause easily.

I am trying from my end and mean while i am also requesting you to compare this code with our HALCoGen generated startup code which doesn't have this issue.

--
Thanks & regards,
Jagadish.

Hi Jagadish, 

I had tried with the HALCoGen code as well, but when I add the memory tests in that it shows the same behaviour.

Thankyou.

Gobind Singh

Hi Jagadish,

Any update on this?

Thank you.

Gobind Singh

Hi Gobind Singh,

The problem is that this code is not TI code, and it is not generated from our HALCoGen or using from our SDL(SafeTI Diagnostic Library).

And this code is written in assembly. So, it will take lot of time to understand it and debug it.

We have our SDL to test memories and peripherals, why don't you use this SDL code for the testing's?

SAFETI_DIAG_LIB Driver or library | TI.com

--
Thanks & regards,
Jagadish.

Hi Jagadish,

I have now used the Halcogen generated code file "HL_sys_core.asm" to initialise the registers. I have now simplified the code to just three function calls, two of which are halcogen created. Third one is just a test writing and reading back from RAM. I have attached the build to replicate the issue.

My problem is, I am unable to debug the issue when it occurs only after the power cycle. By the time I re-connect the emulator, the problem has already occurred. And I am not even using a custom board, it is a TI evaluation board. And now the code is also Halcogen generated.

Thank you.

Gobind Singh

02_StandaloneBoot.zip

Hi Gobind Singh,

My apologies i was off for few days due to festival.

Can you please refer below threads once:

(8) TMS570 RAM state after nPORRST - Arm-based microcontrollers forum - Arm-based microcontrollers - TI E2E support forums

(8) RM57L843: EXTRST and PORST bits in SYSESR register - Arm-based microcontrollers forum - Arm-based microcontrollers - TI E2E support forums

I think power off and PORRST have different behavior effect on RAM, as mentioned in above threads even for PORRST also the contents of the RAM will not get effect and only at on real power off condition the RAM will contain garbage values.

I am suspecting due to this behavior only we are getting RAM test failure.

Can you please also try to include memInit before performing RAM testings?

--
Thanks & regards,
Jagadish.