TMS570LC4357: TMS570LC4357: Example code for F021 OTP Programming

Hi Darryl,

It is similar to the Bank writing only, but the difference is you no need to do any prior initializations or erase operations before writing. 

I never did but on my understanding, you can just give the OTP region address and data directly to the "Fapi_issueProgrammingCommand".

But while writing into the OTP region you have to make sure the following things:

1. If you write to program the customer OTP, make sure you do not unintentionally write to the AJSM (Advanced JTAG Security Module) locations (0xF0000000-0xF000000F), because this will lock the part. You can write data starting at address 0xF0000010.

1. Do I need Fapi_initializeFlahsBanks(SYS_CLK_FREQ) ?

2. Do I need Fapi_setActiveFlashBank(<bank>)  and if so, what bank would that be?

3. Do I need to execute this code from RAM or from a specific Flash bank?

This is why an example would be so helpful.

Hi Darryl,

As i said, i never did this before. Give me time i will try to create one example for customer OTP write.

--
Thanks & regards,
Jagadish.

Hi Darryl,

Using below thread i was created one example project:

(+) CCS/RM46L852: OTP programming - Arm-based microcontrollers forum - Arm-based microcontrollers - TI E2E support forums

And i was successfully able to write into the customer OTP, here is my project:

Customer_OTP_Write_Test_LC4357.zip

--
Thanks & regards,
Jagadish.

1. According to your project's HalCoGen files, you are running at 300MHz, but SYS_CLK_FREQ is set to 150.  Is there a problem with writing too fast to the OTP?

2. My code is running at 150MHz, if i use your example code, it hangs in the Fapi_InitializeFlashBanks(150).  If I instead specify 75MHz, the write to OTP works but the code hangs waiting for Fapi_Status_FsmBusy.  Is there some other HalCoGen setting that mattes tor OTP?

Hi Darryl,

Darryl Uchitil said:

1. According to your project's HalCoGen files, you are running at 300MHz, but SYS_CLK_FREQ is set to 150.  Is there a problem with writing too fast to the OTP?

The argument to the Fapi_initializeFlashBanks function is HCLK frequency not GCLK frequency.

That is the reason i passed the 150

Darryl Uchitil said:

2. My code is running at 150MHz, if i use your example code, it hangs in the Fapi_InitializeFlashBanks(150).  If I instead specify 75MHz, the write to OTP works but the code hangs waiting for Fapi_Status_FsmBusy.  Is there some other HalCoGen setting that mattes tor OTP?

You should pass the HCLK frequency.

--
Thanks & regards,
Jagadish.

To re-iterate #2, My code is running at 150MHz, if I specify 75MHz (The HCLK frequency), the write to OTP works but the code hangs.  Actually ends up in undefined instruction trap, but the OTP was written.  The flash programming code is running from RAM using these functions from another example

_copyAPI2RAM_(&apiLoadStart, &apiRunStart, &apiLoadSize);
_copyAPI2RAM_(&constLoadStart, &constRunStart, &constLoadSize);

I am able to program the code flash in Bank0 and Bank1 and EEPROM section with no issues.  The MPU is not enabled.  Is there something else I need to do to access OTP?

If I single-step through the Fapi_issueProgrammingCommand function in RAM it succeeds and returns successfully.  If I let it run, I get illegal instruction trap. This seems like it could be a cache issue but the cache is not enabled.

Hi Darryl,

You are right it is working if we did step by step debugging and while running continuous it ends up in "prefetchEntry".

I tried different methods changing MPU settings but none of them helpful and no idea why this is happening.

Let me discuss with my senior collogue once and will provide an update on this.

--

Thanks & regards,
Jagadish.

Hi Darryl,

Still didn't get chance to discuss and i will try to discuss ASAP.

--

Thanks & regards,
Jagadish.

Hi Darryl,

I discussed with my senior collogue and even he didn't have any idea why this issue is occurring.

--

Thanks & regards,
Jagadish.

Jagadish,

    I am also having a similar problem with the safety library where the RAM code for testing the Flash, SL_SelfTest_Flash  seems to work with the debugger but fails (or crashes) when run from cold reset without the debugger.  Could this be a related problem?  SL_SelfTest_Flash runs from inside HL_sys_startup, no MPU or cache.

Hi Darryl,

Can you please provide a simplest project to trigger the issue at my end?

Can you please also provide some screenshots for the issue you are talking about?

--

Thanks & regards,
Jagadish.