| Message ID | 20190502025659.30351-1-xiyou.wangcong@gmail.com |
|---|---|
| State | Accepted |
| Delegated to: | David Miller |
| Headers | show |
| Series | [net-next,v2] net: add a generic tracepoint for TX queue timeout | expand |
Thu, May 02, 2019 at 04:56:59AM CEST, xiyou.wangcong@gmail.com wrote: >Although devlink health report does a nice job on reporting TX >timeout and other NIC errors, unfortunately it requires drivers >to support it but currently only mlx5 has implemented it. >Before other drivers could catch up, it is useful to have a >generic tracepoint to monitor this kind of TX timeout. We have >been suffering TX timeout with different drivers, we plan to >start to monitor it with rasdaemon which just needs a new tracepoint. > >Sample output: > > ksoftirqd/1-16 [001] ..s2 144.043173: net_dev_xmit_timeout: dev=ens3 driver=e1000 queue=0 > >Cc: Eran Ben Elisha <eranbe@mellanox.com> >Cc: Jiri Pirko <jiri@mellanox.com> >Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com> Useful. Thanks! Acked-by: Jiri Pirko <jiri@mellanox.com>
On 5/2/2019 5:56 AM, Cong Wang wrote: > Although devlink health report does a nice job on reporting TX > timeout and other NIC errors, unfortunately it requires drivers > to support it but currently only mlx5 has implemented it. > Before other drivers could catch up, it is useful to have a > generic tracepoint to monitor this kind of TX timeout. We have > been suffering TX timeout with different drivers, we plan to > start to monitor it with rasdaemon which just needs a new tracepoint. > > Sample output: > > ksoftirqd/1-16 [001] ..s2 144.043173: net_dev_xmit_timeout: dev=ens3 driver=e1000 queue=0 > > Cc: Eran Ben Elisha <eranbe@mellanox.com> > Cc: Jiri Pirko <jiri@mellanox.com> > Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com> Reviewed-by: Eran Ben Elisha <eranbe@mellanox.com>
From: Cong Wang <xiyou.wangcong@gmail.com> Date: Wed, 1 May 2019 19:56:59 -0700 > Although devlink health report does a nice job on reporting TX > timeout and other NIC errors, unfortunately it requires drivers > to support it but currently only mlx5 has implemented it. > Before other drivers could catch up, it is useful to have a > generic tracepoint to monitor this kind of TX timeout. We have > been suffering TX timeout with different drivers, we plan to > start to monitor it with rasdaemon which just needs a new tracepoint. > > Sample output: > > ksoftirqd/1-16 [001] ..s2 144.043173: net_dev_xmit_timeout: dev=ens3 driver=e1000 queue=0 > > Cc: Eran Ben Elisha <eranbe@mellanox.com> > Cc: Jiri Pirko <jiri@mellanox.com> > Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com> Applied, thanks Cong.
diff --git a/include/trace/events/net.h b/include/trace/events/net.h index 1efd7d9b25fe..2399073c3afc 100644 --- a/include/trace/events/net.h +++ b/include/trace/events/net.h @@ -95,6 +95,29 @@ TRACE_EVENT(net_dev_xmit, __get_str(name), __entry->skbaddr, __entry->len, __entry->rc) ); +TRACE_EVENT(net_dev_xmit_timeout, + + TP_PROTO(struct net_device *dev, + int queue_index), + + TP_ARGS(dev, queue_index), + + TP_STRUCT__entry( + __string( name, dev->name ) + __string( driver, netdev_drivername(dev)) + __field( int, queue_index ) + ), + + TP_fast_assign( + __assign_str(name, dev->name); + __assign_str(driver, netdev_drivername(dev)); + __entry->queue_index = queue_index; + ), + + TP_printk("dev=%s driver=%s queue=%d", + __get_str(name), __get_str(driver), __entry->queue_index) +); + DECLARE_EVENT_CLASS(net_dev_template, TP_PROTO(struct sk_buff *skb), diff --git a/net/sched/sch_generic.c b/net/sched/sch_generic.c index 848aab3693bd..cce1e9ee85af 100644 --- a/net/sched/sch_generic.c +++ b/net/sched/sch_generic.c @@ -32,6 +32,7 @@ #include <net/pkt_sched.h> #include <net/dst.h> #include <trace/events/qdisc.h> +#include <trace/events/net.h> #include <net/xfrm.h> /* Qdisc to use by default */ @@ -441,6 +442,7 @@ static void dev_watchdog(struct timer_list *t) } if (some_queue_timedout) { + trace_net_dev_xmit_timeout(dev, i); WARN_ONCE(1, KERN_INFO "NETDEV WATCHDOG: %s (%s): transmit queue %u timed out\n", dev->name, netdev_drivername(dev), i); dev->netdev_ops->ndo_tx_timeout(dev);
Although devlink health report does a nice job on reporting TX timeout and other NIC errors, unfortunately it requires drivers to support it but currently only mlx5 has implemented it. Before other drivers could catch up, it is useful to have a generic tracepoint to monitor this kind of TX timeout. We have been suffering TX timeout with different drivers, we plan to start to monitor it with rasdaemon which just needs a new tracepoint. Sample output: ksoftirqd/1-16 [001] ..s2 144.043173: net_dev_xmit_timeout: dev=ens3 driver=e1000 queue=0 Cc: Eran Ben Elisha <eranbe@mellanox.com> Cc: Jiri Pirko <jiri@mellanox.com> Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com> --- include/trace/events/net.h | 23 +++++++++++++++++++++++ net/sched/sch_generic.c | 2 ++ 2 files changed, 25 insertions(+)