Re: [PATCH net-next v04 1/4] hinic3: Fix the RX stalls in hinic3_rx_poll()
From: Jakub Kicinski
Date: Wed Nov 18 2026 - 21:09:19 EST
On Tue, 17 Nov 2026 20:08:31 +0800 Fan Gong wrote:
> hinic3_rx_poll() can leave the loop with a completed CQE still sitting at
> the head of the ring, or with no WQE left to receive into, and still
> return nr_pkts, which is below budget. hinic3_poll() then calls
> napi_complete_done() and re-enables the MSI-X interrupt, which is edge
> triggered, so nothing re-arms the queue if no further packet arrives.
Look how you Hauwei coworker is fixing this in hns3:
https://lore.kernel.org/all/20261118130330.3026625-4-shaojijie@xxxxxxxxxx/
That's much better. Live-locking the system in NAPI when its OOM
is pretty bad. My personal preference is to have a work that just
checks for Rx buffer queue being empty every few sec.