[Openocd-development] Patch for Luminary Problems

[Michael S. <rin...@di...>]

Hi,

I did some digging - it seems that the write to NVIC_AIRCR in cortex_m3_assert_reset causes the error:

>value captured during scan didn't pass the requested check: captured: 0x00 check_value: 0x01 check_mask: 0x0f
>in_handler reported a failed check
>Invalid ACK in SWJDP transaction

I still have no solution for that problem. However, with the attached patch, I can now "reset halt" the CPU, the target stops at the reset vector, *and* resume/step work fine without causing a NMI.

It seems any access to NVIC_AIRCR works to fix the problem - either a read, or a write with only the key value but no reset bits set. This looks like some synchronization problem to me, but I will leave that to someone who knows more about the details. I chose a read, since that should have least impact on previously working targets.

cu
Michael