On Thu, 17 Nov 2016 05:20:50 -0800 Eric Dumazet <eric.duma...@gmail.com> wrote:
> On Thu, 2016-11-17 at 09:16 +0100, Jesper Dangaard Brouer wrote: > > > > > I noticed there is a Send-Q, and the perf-top2 is _raw_spin_lock, which > > looks like it comes from __dev_queue_xmit(), but we know from > > experience that this stall is actually caused by writing the > > tailptr/doorbell in the HW. Thus, this could benefit a lot from > > bulk/xmit_more into the qdisc layer. > > The Send-Q is there because of TX-completions being delayed a bit, > because of IRQ mitigation. > > (ethtool -c eth0) > > It happens even if you do not have a qdisc in the first place. > > And we do have xmit_more in the qdisc layer already. I can see that qdisc layer does not activate xmit_more in this case. -- Best regards, Jesper Dangaard Brouer MSc.CS, Principal Kernel Engineer at Red Hat Author of http://www.iptv-analyzer.org LinkedIn: http://www.linkedin.com/in/brouer $ ethtool -c mlx5p4 Coalesce parameters for mlx5p4: Adaptive RX: on TX: off stats-block-usecs: 0 sample-interval: 0 pkt-rate-low: 0 pkt-rate-high: 0 rx-usecs: 3 rx-frames: 32 rx-usecs-irq: 0 rx-frames-irq: 0 tx-usecs: 16 tx-frames: 32 tx-usecs-irq: 0 tx-frames-irq: 0 rx-usecs-low: 0 rx-frame-low: 0 tx-usecs-low: 0 tx-frame-low: 0 rx-usecs-high: 0 rx-frame-high: 0 tx-usecs-high: 0 tx-frame-high: 0