Question about S32k312 C40_Ip write/erase pflash

[lucaswang]

hello!

I'm using s32k312 with RTD5.0.0, there's some problem about C40_Ip erasing pflash :

1. Erase block0(0x420000-0x500000) failed, will cause hard_fault;
2. Erase block1(0x500000-Hse_start) successful;

I found a post https://community.nxp.com/t5/S32K/S32K344-C40-IP-Hardware-Fault-Problem/td-p/1697432 

I modify C40_Ip.h, but the erasing still fails.

[danielmartynek]

Hi @lucaswang,

It works with this test project:

https://community.nxp.com/t5/S32K-Knowledge-Base/S32K312-C40-Ip-SRAM-RTD-500-DS35/ta-p/2074245

Make sure these functions are also placed in SRAM (SchM_Mem_43_INFLS.h).

extern void SchM_Enter_Mem_43_INFLS_MEM_EXCLUSIVE_AREA_xx(void);
extern void SchM_Exit_Mem_43_INFLS_MEM_EXCLUSIVE_AREA_xx(void);

Regards,

Daniel

[lucaswang]

Thank you for your reply! According to what you said, I can perform the erase and write operations. However, I still encountered other problems. My program receives data sent from the PC via the serial port and writes to the FLASH while receiving the data. But each time, it can only write a part of the data, and then the program triggers a HARDFAULT reset. Here is the interface of my write function:

[danielmartynek]

Hello @lucaswang,

There can be a collision if 

• the Interrupt vector table
• or the interrupt handlers
• or the data that the handlers read

are in that PFlash block that is being programmed.

Can you mask the interrupts before the flash operations are launched?

Regards,

Daniel

[lucaswang]

I have masked the interrupts before the flash operations.

See my function interface：

I found that when the length is passed in as 128, the writing can be completed. However, when the length is passed in as 4096, only a part of it can be written, and then a reset will be triggered afterwards.

I have another question, as shown in the following picture：

 this sector is locked. How should I unlock it? I'm using a JFlash debugger, and after the problem occurred, the program can no longer be flashed in.

[danielmartynek]

This is expected, as you can see in the RM, we can program max. 128B at once.

Regards,

Daniel