Daniel Leung
Daniel Leung
We are going in the other direction so this is no longer needed.
I support @teburd's suggestion to flush logging. For logging to NOT copy the format strings and parameters into the buffer, there are specific conditions that need to be met first....
It would be a good idea to separate the changes from the first commit into: arch, subsys and SoC. The backend changes should be squashed into one commit.
Could you move the changes under `tests/` from first commit to the last commit?
The original design is to let the application decide what to do in this situation by implementing `k_sys_fatal_error_handler()`. The default implementation is simply to halt the system, but the application...
> > The original design is to let the application decide what to do in this situation by implementing `k_sys_fatal_error_handler()`. The default implementation is simply to halt the system, but...
From my limited debugging, it seems like on PTL, the busy bit is set after sending the first message, which prevents any more messages to be sent.
I tested it a bit more. It seems like the exception comes when requesting D0->D3 transition more than one time. The test showed that the first D0->D3->D0 worked fine and...
The issue with D0->D3->D0 is fixed so suspend/resume should be working on LNL and PTL.
Zephyr side PR has been merged.