A follow-up on this patch: I am pausing it in its current form. The exact payload-length checks here were based on my earlier ring-buffer change, and Steven has rejected that direction because the ring buffer is meant to keep the allocated size, not the requested payload size.
So this selftest should not move forward as-is. If there is a real fix, it needs to come from an event-specific change that explicitly stores the true payload length where needed. I will revisit it only from that direction. Thanks, Cao Ruichuang
