svc: Report xprt dequeue latency

Record the time between when a rqstp is enqueued on a transport
and when it is dequeued. This includes how long the rqstp waits on
the queue and how long it takes the kernel scheduler to wake a
nfsd thread to service it.

The svc_xprt_dequeue trace point is altered to include the number
of microseconds between xprt_enqueue and xprt_dequeue.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: J. Bruce Fields <bfields@redhat.com>
This commit is contained in:
Chuck Lever 2018-03-27 10:52:27 -04:00 коммит произвёл J. Bruce Fields
Родитель aaba72cd4e
Коммит 55f5088c22
3 изменённых файлов: 29 добавлений и 6 удалений

Просмотреть файл

@ -272,6 +272,7 @@ struct svc_rqst {
#define RQ_BUSY (6) /* request is busy */
#define RQ_DATA (7) /* request has data */
unsigned long rq_flags; /* flags field */
ktime_t rq_qtime; /* enqueue time */
void * rq_argp; /* decoded arguments */
void * rq_resp; /* xdr'd results */

Просмотреть файл

@ -657,14 +657,36 @@ DECLARE_EVENT_CLASS(svc_xprt_event,
show_svc_xprt_flags(__entry->flags))
);
DEFINE_EVENT(svc_xprt_event, svc_xprt_dequeue,
TP_PROTO(struct svc_xprt *xprt),
TP_ARGS(xprt));
DEFINE_EVENT(svc_xprt_event, svc_xprt_no_write_space,
TP_PROTO(struct svc_xprt *xprt),
TP_ARGS(xprt));
TRACE_EVENT(svc_xprt_dequeue,
TP_PROTO(struct svc_rqst *rqst),
TP_ARGS(rqst),
TP_STRUCT__entry(
__field(struct svc_xprt *, xprt)
__field(unsigned long, flags)
__field(unsigned long, wakeup)
__string(addr, rqst->rq_xprt->xpt_remotebuf)
),
TP_fast_assign(
__entry->xprt = rqst->rq_xprt;
__entry->flags = rqst->rq_xprt->xpt_flags;
__entry->wakeup = ktime_to_us(ktime_sub(ktime_get(),
rqst->rq_qtime));
__assign_str(addr, rqst->rq_xprt->xpt_remotebuf);
),
TP_printk("xprt=%p addr=%s flags=%s wakeup-us=%lu",
__entry->xprt, __get_str(addr),
show_svc_xprt_flags(__entry->flags),
__entry->wakeup)
);
TRACE_EVENT(svc_wake_up,
TP_PROTO(int pid),

Просмотреть файл

@ -409,6 +409,7 @@ void svc_xprt_do_enqueue(struct svc_xprt *xprt)
if (test_and_set_bit(RQ_BUSY, &rqstp->rq_flags))
continue;
atomic_long_inc(&pool->sp_stats.threads_woken);
rqstp->rq_qtime = ktime_get();
wake_up_process(rqstp->rq_task);
goto out_unlock;
}
@ -530,7 +531,6 @@ void svc_wake_up(struct svc_serv *serv)
if (test_bit(RQ_BUSY, &rqstp->rq_flags))
continue;
rcu_read_unlock();
dprintk("svc: daemon %p woken up.\n", rqstp);
wake_up_process(rqstp->rq_task);
trace_svc_wake_up(rqstp->rq_task->pid);
return;
@ -726,7 +726,7 @@ out_found:
rqstp->rq_chandle.thread_wait = 5*HZ;
else
rqstp->rq_chandle.thread_wait = 1*HZ;
trace_svc_xprt_dequeue(rqstp->rq_xprt);
trace_svc_xprt_dequeue(rqstp);
return rqstp->rq_xprt;
}