[SCSI] zfcp: optimize zfcp_qdio_account

Remove expensive ktime_get()/ktime_us_delta() functions from the hot
path and use get_clock_monotonic() instead. This elimates seven
function calls and avoids a lot of unnecessary calculations.

Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Christof Schmitt <christof.schmitt@de.ibm.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

drivers/s390/scsi/zfcp_def.h

@@ -438,7 +438,7 @@ struct zfcp_qdio {
 	struct zfcp_qdio_queue	req_q;
 	spinlock_t		stat_lock;
 	spinlock_t		req_q_lock;
-	ktime_t			req_q_time;
+	unsigned long long	req_q_time;
 	u64			req_q_util;
 	atomic_t		req_q_full;
 	wait_queue_head_t	req_q_wq;

drivers/s390/scsi/zfcp_qdio.c

@@ -56,16 +56,15 @@ static void zfcp_qdio_zero_sbals(struct qdio_buffer *sbal[], int first, int cnt)
 }
 
 /* this needs to be called prior to updating the queue fill level */
-static void zfcp_qdio_account(struct zfcp_qdio *qdio)
+static inline void zfcp_qdio_account(struct zfcp_qdio *qdio)
 {
-	ktime_t now;
-	s64 span;
+	unsigned long long now, span;
 	int free, used;
 
 	spin_lock(&qdio->stat_lock);
-	now = ktime_get();
-	span = ktime_us_delta(now, qdio->req_q_time);
-	free = max(0, atomic_read(&qdio->req_q.count));
+	now = get_clock_monotonic();
+	span = (now - qdio->req_q_time) >> 12;
+	free = atomic_read(&qdio->req_q.count);
 	used = QDIO_MAX_BUFFERS_PER_Q - free;
 	qdio->req_q_util += used * span;
 	qdio->req_q_time = now;