net: add a generic tracepoint for TX queue timeout
authorCong Wang <xiyou.wangcong@gmail.com>
Thu, 2 May 2019 02:56:59 +0000 (19:56 -0700)
committerDavid S. Miller <davem@davemloft.net>
Sat, 4 May 2019 04:41:41 +0000 (00:41 -0400)
commit141b6b2ad75d92770240de3af98d55c41ce7cd18
treec6b302bf24d04bd8a3d294d2df858f60d15116e2
parentf3f050a4df355c398f70e3788360a1262ac0c5df
net: add a generic tracepoint for TX queue timeout

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>
Acked-by: Jiri Pirko <jiri@mellanox.com>
Reviewed-by: Eran Ben Elisha <eranbe@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
include/trace/events/net.h
net/sched/sch_generic.c