Hello Rusty, What's the reason to use refill work in receiving path? static void refill_work(struct work_struct *work) { struct virtnet_info *vi; bool still_empty; vi = container_of(work, struct virtnet_info, refill.work); napi_disable(&vi->napi); still_empty = !try_fill_recv(vi, GFP_KERNEL); napi_enable(&vi->napi); /* In theory, this can happen: if we don't get any buffers in * we will *never* try to fill again. */ if (still_empty) schedule_delayed_work(&vi->refill, HZ/2); } It looks more expensive than only refilling the buffers in recv path. It could cause more guest exits, maybe more RX interrupts. Can we move the refill work in recv path only? If we can remove it, I will run some test to compare the difference with the change. Thanks Shirley -- To unsubscribe from this list: send the line "unsubscribe kvm" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html