SBSFU | STM32F756ZGTx | Issue with running a port

[Jakub Standarski]

Hello all.
I have troubles porting SBSFU onto STM32F756ZGTx.

I took 1 image sample for STM32F769I as a base. I modified:
- STM32F756ZGTx_FLASH.ld (SECoreBin)
- STM32F756ZGTx_FLASH.ld (SBSFU)
- mapping_sbsfu.ld
- mapping_fwimg.ld
- Other changes to source code under SBSFU and SECoreBin to work with target platform.

Unfortunately, it still doesn't work and I'm not able to find out why.
I connected debugger and every time programs ends up in a hard fault handler...
With SFU_MPU_PROTECT_ENABLE turned off:

With SFU_MPU_PROTECT_ENABLE turned on:
First debugger starts here

Then it fails upon calling SFU_BOOT_SystemSecurity_Config() -> SFU_COM_Init() -> setvbuf():

 

If necessary I can provide snippets of code. Thanks!

 

 

 

[RhSilicon]

I don't know if it can help in your case, but I found these videos about advanced debugging:

[STM32CubeIDE Advanced Debug Features]

Complement for some models:

[Dual Core Debugging on STM32H7 with STM32CubeIDE]

[Jakub Standarski]

Thanks! I didn't solve the issue but did help finding this out:

 

That's the information from built in build analyzer for SBSFU.elf. I have no clue why SE_Entry_Secure_ROM_Region is that huge occupying 3729% of allocated memory.

[RhSilicon]

---

3729%

---

Went a little over the top :grinning_face_with_sweat:

[Jakub Standarski]

Just a little bit :D
I would like to know why since we are stuck on it for a long time. I saw similar thread on that forum but nobody replied to it.

[RhSilicon]

There are some videos about SBSFU on YouTube:

https://www.youtube.com/@stmicroelectronics/search?query=SBSFU

[Jakub Standarski]

Yeah, I know. But doesn't help too much. Maybe I have to watch them again and find that missing piece. I hoped that someone experienced something similar and knows how to deal with it.

[Jakub Standarski]

Okay, so that's weird, but I built two different sample projects:

1. NUCLEO-L432KC (1_Image)
   

    

2. STM32F413H-Discovery (1_Image)
   

Both presents same misleading information that there is some issue with memory mapping. The thing is that NUCLEO-L432KC works totally fine. Looks like it's not an issue with the build itself, but with the tool or naming used throughout linker files. Correct me if I'm wrong.

[Jakub Standarski]

Problem solved. Issue came from missing line inside SBSFU linker script. So next time I should be more focused :)

PS. It doesn't mean that tool or naming is fine. Build analyzer shouldn't show misleading information.