When we are within hard_start_xmit, the queue is already awake.
Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
        if (ether1394_send_packet(ptask, tx_len))
                goto fail;
 
-       netif_wake_queue(dev);
        return NETDEV_TX_OK;
 fail:
        if (ptask)
        priv->stats.tx_errors++;
        spin_unlock_irqrestore(&priv->lock, flags);
 
-       if (netif_queue_stopped(dev))
-               netif_wake_queue(dev);
-
        /*
         * FIXME: According to a patch from 2003-02-26, "returning non-zero
         * causes serious problems" here, allegedly.  Before that patch,