TMDSCNCD263: AM263x: sbl_qspi_am263x example stucks during HSM's runtime firmware load

Hi vanni vinti,

Did you try using a default SBL_QSPI example with this ? Does that work for you ?

Best Regards,

Aakash

Hi,

If by default SBL_QSPI example you mean the precompiled executable available in SDK, the answer is yes and in that case, the HSM firmware loading seems to work correctly. In fact, by programming the SBL binary in flash and then running the "hsm_service..." sample project, I am able to read correctly the version of the HSM loaded by the SBL.

BR,

Vanni

Hi vanni vinti,

So it means your hardware is perfectly alright to run the HSM firmware. There is some problem to run the SBL. Can you share more changes w.r.t sbl-qspi that has been run ?

vanni vinti said:

the code stucks during the HSM firmware's load

Also, how are you sure that the software is stuck at the mentioned place ? Can you try this method to debug SBL - https://e2e.ti.com/support/microcontrollers/arm-based-microcontrollers-group/arm-based-microcontrollers/f/arm-based-microcontrollers-forum/1119575/faq-mcu-plus-sdk-am243x-how-to-debug-sbl_qspi-for-am263-and-am273-or-sbl_ospi-for-am243-or-an-application-which-is-booted-via-ospi-qspi-boot

Do load the symbols and tell me the exact location of your program counter ? Does it matches with SBL symbols ?

Best Regards,
Aakash

Hi,

in my previous tests I just downloaded via JTAG the code in RAM and run it. I noticed in this way that the code was stucking inside an internal TI's API waiting for una interrupt coming from HSM. Consider that also flashing the code and running it, it seems to stuck, since there is no jump to an already flashed application code.

By the way I will try to flash it again and then debug it in attached mode as you seggested. I will let you know.

BR,

Vanni

Hi,

I did the test you suggested:

This is when I start to debug in attach mode. I can correctly read symbols (loop variable = 1), as you can see.

This is the failing function:

and the following is the error cause ( Bootloader_socLoadHsmRtFw function):

Please note that the code is exactly the one provided in the SDK.

Note also that, after the load, the HSM returns a valid CRC but a wrong load status.

Any suggestions about this fault?

Sorry, I did a mistake with the first image that should be the following:

Hi vanni vinti,

Can you check if there is no change in the HSMRtImg.h file from the delivered MCU_PLUS_SDK 09.00 or 09.01 ?

Best Regards,
Aakash

Hi,

this is the file I linked in my project. 

hsmRtImg.h

Please, can you confirm that the sbl_qspi example is working on your environment?

Is there any released 09.01 SDK? I'm not able to find it on TI download page.

BR,

Vanni  

Hi Aakash, can you confirm that they can switch the EVMs to SE enforced mode without risking to brick the units?

Thanks,

Eyal

Hi Aakash, please can you confirm that the operation to move the EVB from HSFA to HSSE mode is completely free from the risk of brikking the EVMs?

Hello Vanni,

There should be no risk of breaking the EVMs by doing this, but if for some reason that did occur, you would be eligible to RMA the EVM to receive a replacement as it would be from TI directed steps that caused the issue so we would provide a replacement.

Best Regards,

Ralph Jacobi