How come the softnet stats total is incremented in both netif_rx() and netif_receive_skb() in net/core/dev.c in the kernel netif_rx() __get_cpu_var(netdev_rx_stat).total++; netif_receive_skb() __get_cpu_var(netdev_rx_stat).total++; It is my understanding that a non-NAPI NIC driver calls netif_rx() to place incoming packets on the backlog queue. The poll method for the backlog queue is process_backlog() which processes the packets. It calls netif_receive_skb() which in turns calls the ingress filter and then passes the packet to its registered protcol handler. In this case it appears to me that the total is incremented twice for a packet on its path up the network stack. Is my understanding of this incorrect ? Jonathan Lynch - : send the line "unsubscribe linux-net" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
