Re: [PATCH net-next V3 3/3] tun: rx batching
From: David Miller <davem@davemloft.net>
Date: 2016-12-31 17:31:47
Also in:
kvm, virtualization
From: Jason Wang <jasowang@redhat.com> Date: Fri, 30 Dec 2016 13:20:51 +0800
quoted hunk ↗ jump to hunk
@@ -1283,10 +1314,15 @@ static ssize_t tun_get_user(struct tun_struct *tun, struct tun_file *tfile, skb_probe_transport_header(skb, 0); rxhash = skb_get_hash(skb); + #ifndef CONFIG_4KSTACKS - local_bh_disable(); - netif_receive_skb(skb); - local_bh_enable(); + if (!rx_batched) { + local_bh_disable(); + netif_receive_skb(skb); + local_bh_enable(); + } else { + tun_rx_batched(tfile, skb, more); + } #else netif_rx_ni(skb); #endif
If rx_batched has been set, and we are talking to clients not using this new MSG_MORE facility (or such clients don't have multiple TX packets to send to you, thus MSG_MORE is often clear), you are doing a lot more work per-packet than the existing code. You take the queue lock, you test state, you splice into a local queue on the stack, then you walk that local stack queue to submit just one SKB to netif_receive_skb(). I think you want to streamline this sequence in such cases so that the cost before and after is similar if not equivalent.