Hi Xuneng,
Thanks for reviewing this patch.
> The diganose of the problem seems good to me. The fix is small and
> helps to mitigate the stall.
Yes, so I think we can start from the existing patch and get it backpatched.
> A more ideal fix seems to be more
> event-driven -- let the late pinners cancel themselves.
If I understand correctly, you are referring to changing the PINBuffer logic
where the backend sees BM_PIN_COUNT_WAITER and cancels itself?
I am not sure if that is the best approach as PinBuffer is in the hot path.
Thanks,
Shihao