HardFault_Handler() error in SPI

[GPaiv.1]

Hi, 

When I leave my code to play for more than one hour I received the error HardFault_Handler(), I found in registers whats the last line executed before entering in the error function, below I put a picture of this part of the code.  

Note: That error is random, don't have a specific moment time when happens.

[DavidAlfa]

Are you changing the SPI pointer at some point? An interrupt in that moment might cause trouble.

Can't help more without seeing any code.

[GPaiv.1]

Hi, Thanks for your answer.

No, I do not change de SPI pointer.

I'm using the demo code ST25R3911B-DISCO, I'm just using mifare_ul.c to read an EEPROM of a Tag.

[Brian TIDAL]

Hi,

hpsi->pRxBuffPtr is at @0x20017fff which is the end of STM32L476 internal SRAM1 area. Access to next byte (@0x20018000) will cause a fault...

it looks like a HAL_SPI_TransmitReceive call with an incorrect/corrupted length parameter...

Rgds

BT

[Tesla DeLorean]

Is the buffer on the stack or heap?

Looks to be blowing the buffer's scope, add some better sanity checking in your code.

Watch for stack/heap overflow, or resource leaks within your implementation.

Remember callbacks are done under interrupt context, and there is likely a lot of non thread-safe code, or other things that are ill-considered.

[GPaiv.1]

Hi,

I resolve the problem increase de size of my buffer then received information by tag.

But now I had a new problem, below has a print of my new problem.

[Brian TIDAL]

Hi,

hspi pointer seems to be corrupted (value 0x44444444: it looks like a buffer overflow has corrupted this pointer....).

I would recommend to check buffers size.

Rgds

BT

[DavidAlfa · Accepted solution]

Are you using dynamic allocation? (malloc() ,calloc()...).

If so, do your best to avoid that, as it's completely broken.

Better to declare an array and use it.

What I do when using big buffers but none at the same time,​ is making a big union with them.

So I can use any without eating up all my RAM.