Reset from CPU lock-up detected

RE: Reset from CPU lock-up detected

Amanda Hsieh — Thu, 19 Oct 2023 13:15:50 GMT

I see you have created a new support case for the latest issue which is away form the original scope. Please use that case to continue.

RE: Reset from CPU lock-up detected

kevin101 — Wed, 18 Oct 2023 20:03:03 GMT

I'm trying to understand the 'timeslots'. This seems to have something do to with the radio. We are not using the radio calls explicitiy, but we are using bluetooth. Does bluetooth use these timeslots? Are you saying that our callbacks could be taking too long?
Is there a way to see which timeslot is taking too long? Are there definitions of the size of the timeslots? Are they adjustable?

RE: Reset from CPU lock-up detected

Amanda Hsieh — Wed, 18 Oct 2023 18:23:32 GMT

Hi,

The assertion 0x0001089E can occur if you have a timeslot that overstayed its duration, or that you are stepping through the code while debugging.

-Amanda H.

RE: Reset from CPU lock-up detected

kevin101 — Wed, 18 Oct 2023 14:42:40 GMT

I think this may be related: We often get a fault from the SoftDevice with ID=1 at PC=0x0001089e.

Is there a way to determine what part of the SoftDevice is posting this fault?

RE: Reset from CPU lock-up detected

Turbo J — Fri, 13 Oct 2023 18:09:00 GMT

The hardfault handler in the softdevice can lock up if the stack pointer is messed up (like not pointing into RAM etc).

Its does more work than just calling the application handler, and this work assumes working stack.

Look through your code for stack buffer overflows.

RE: Reset from CPU lock-up detected

kevin101 — Fri, 13 Oct 2023 17:55:44 GMT

From the answer in the linked question:

"Hi, you should possible check your hardfault handler, for instance a hardfault within the hardfault handler may cause reset by CPU lockup."

Are there other causes of the CPU lock-up?

We did not have a HardFault Handler implemented so we were using the default from 'gcc_startup_nrf52.S' which looks like it just sits in and endless loop.

We ended up implementing a HardFault Handler that just saves some information to a section of RAM that does not get initialized at boot. This did not seem to make any difference, and it doesn't look like the HardFault handler is called in the case of the CPU lock-up. (I am able to force a HardFault so I verified that our handler is working.)

RE: Reset from CPU lock-up detected

Amanda Hsieh — Tue, 26 Sep 2023 13:24:02 GMT

Hi,

See this answer.

Regards,
Amanda H.