[net-next PATCH 4/5] pktgen: avoid expensive set_current_state() call in loop

[Jesper Dangaard Brouer <brouer@redhat.com>]

I request review as I'm uncertain of this change as I don't know
the API of set_current_state() very well.

The set_current_state(TASK_INTERRUPTIBLE) uses a xchg, which implicit
is LOCK prefixed.

Avoid calling set_current_state() inside the busy-loop in
pktgen_thread_worker().  In case of pkt_dev->delay, then it is still
used in pktgen_xmit() via the spin() call.

Performance data with CLONE_SKB==100000 and TX ring buffer size=1024:
 (single CPU performance, ixgbe 10Gbit/s, E5-2630)
 * Prev:  5608781 pps --> 178.29ns (1/5608781*10^9)
 * Now:   5857065 pps --> 170.73ns (1/5857065*10^9)
 * Diff:  +248284 pps -->  -7.56ns

Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
---

 net/core/pktgen.c |    9 +++------
 1 files changed, 3 insertions(+), 6 deletions(-)

diff --git a/net/core/pktgen.c b/net/core/pktgen.c
index 7752806..cae7e0c 100644
--- a/net/core/pktgen.c
+++ b/net/core/pktgen.c
@@ -3409,10 +3409,10 @@ static int pktgen_thread_worker(void *arg)
 
 	pr_debug("starting pktgen/%d:  pid=%d\n", cpu, task_pid_nr(current));
 
-	set_current_state(TASK_INTERRUPTIBLE);
-
 	set_freezable();
 
+	__set_current_state(TASK_RUNNING);
+
 	while (!kthread_should_stop()) {
 		pkt_dev = next_to_run(t);
 
@@ -3426,8 +3426,6 @@ static int pktgen_thread_worker(void *arg)
 			continue;
 		}
 
-		__set_current_state(TASK_RUNNING);
-
 		if (likely(pkt_dev)) {
 			pktgen_xmit(pkt_dev);
 
@@ -3458,9 +3456,8 @@ static int pktgen_thread_worker(void *arg)
 		}
 
 		try_to_freeze();
-
-		set_current_state(TASK_INTERRUPTIBLE);
 	}
+	set_current_state(TASK_INTERRUPTIBLE);
 
 	pr_debug("%s stopping all device\n", t->tsk->comm);
 	pktgen_stop(t);