irda: irda-usb: Replace timeval with ktime_t

The irda usb driver uses 'timeval', which we try to remove in the kernel
because all 32-bit time types will break in the year 2038.

This patch also changes do_gettimeofday() to ktime_get() accordingly,
since ktime_get returns a ktime_t, but do_gettimeofday returns a
struct timeval, and the other reason is that ktime_get() uses
the monotonic clock.

This patch uses ktime_us_delta to get the elapsed time, and in this
way it no longer needs to check for the overflow, because
ktime_us_delta returns time difference of microsecond.

Signed-off-by: Chunyan Zhang <zhang.chunyan@linaro.org>
Reviewed-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: David S. Miller <davem@davemloft.net>

drivers/net/irda/irda-usb.c

@@ -495,18 +495,12 @@ static netdev_tx_t irda_usb_hard_xmit(struct sk_buff *skb,
 		mtt = irda_get_mtt(skb);
 		if (mtt) {
 			int diff;
-			do_gettimeofday(&self->now);
-			diff = self->now.tv_usec - self->stamp.tv_usec;
+			diff = ktime_us_delta(ktime_get(), self->stamp);
 #ifdef IU_USB_MIN_RTT
 			/* Factor in USB delays -> Get rid of udelay() that
 			 * would be lost in the noise - Jean II */
 			diff += IU_USB_MIN_RTT;
 #endif /* IU_USB_MIN_RTT */
-			/* If the usec counter did wraparound, the diff will
-			 * go negative (tv_usec is a long), so we need to
-			 * correct it by one second. Jean II */
-			if (diff < 0)
-				diff += 1000000;
 
 		        /* Check if the mtt is larger than the time we have
 			 * already used by all the protocol processing
@@ -869,7 +863,7 @@ static void irda_usb_receive(struct urb *urb)
 	 * reduce the min turn time a bit since we will know
 	 * how much time we have used for protocol processing
 	 */
-        do_gettimeofday(&self->stamp);
+	self->stamp = ktime_get();
 
 	/* Check if we need to copy the data to a new skb or not.
 	 * For most frames, we use ZeroCopy and pass the already