Hi all,
We are testing the failure modes for the OCTOSPI memory mapped mode and there isn't much info in the documentation.
The setup:
* STM32L4S9ZIJ6
* Dual external RAM chips (CY15B104QSN) configured for dual QPI DTR.
* External memory used for TouchGFX driven display framebuffers (might be used for alternate purposes in the future).
To simulate the failure of the external RAM chip we are reconfiguring the pins to GPIO operation after the initial configuration has finished and the device is operating normally.
We are seeing the following symptoms after introducing the error.
Error on data or CS lines
CPU continues normal read/write operation with incorrect data reads. As far as we can see the QPI protocol for reads and writes is not able to even detect that the external device is present.
Is there any way to check at runtime when this happens? Only solution we have come up is periodically reading/writing known data to check the operation.
Error on the CLK line
There seem to be two very different behaviours for this:
* External RAM is accessed by the CPU. This causes an immediate CPU lockup (debugger fails, debugger cannot reattach, hard fault is not triggered, no indication that the CPU is still running).
* External RAM is accessed by the hardware users (DMA2D and LTDC). On both cases the transactions seem to fail (no interrupts) and LTDC is frequently reporting FIFO underrun.
Is there any way to detect the HW failure for DMA2D and LTDC?
Is there any documentation regarding CPU the lockup or a way to recover?
Cheers
