[PATCH 13/39]: pkt_sched: Add qdisc_all_tx_empty()

[PATCH 13/39]: pkt_sched: Add qdisc_all_tx_empty()

[David Miller]

As per comments, this isn't a foolproof implementation.  This is being
added so that we can contain and isolate all the explicit ->tx_queue
references in the tree.

Signed-off-by: David S. Miller <davem@davemloft.net>
---
 include/net/irda/irda_device.h |    2 +-
 include/net/sch_generic.h      |   13 +++++++++++++
 2 files changed, 14 insertions(+), 1 deletions(-)

diff --git a/include/net/irda/irda_device.h b/include/net/irda/irda_device.h
index 16fbf67..3025ae1 100644
--- a/include/net/irda/irda_device.h
+++ b/include/net/irda/irda_device.h
@@ -223,7 +223,7 @@ int  irda_device_is_receiving(struct net_device *dev);
 /* Interface for internal use */
 static inline int irda_device_txqueue_empty(const struct net_device *dev)
 {
-	return skb_queue_empty(&dev->tx_queue.qdisc->q);
+	return qdisc_all_tx_empty(dev);
 }
 int  irda_device_set_raw_mode(struct net_device* self, int status);
 struct net_device *alloc_irdadev(int sizeof_priv);
diff --git a/include/net/sch_generic.h b/include/net/sch_generic.h
index a71994b..f83e860 100644
--- a/include/net/sch_generic.h
+++ b/include/net/sch_generic.h
@@ -191,6 +191,19 @@ static inline void qdisc_reset_all_tx(struct net_device *dev)
 	qdisc_reset(dev->tx_queue.qdisc);
 }
 
+/* Are all TX queues of the device empty?  */
+static inline bool qdisc_all_tx_empty(const struct net_device *dev)
+{
+	struct netdev_queue *txq = &dev->tx_queue;
+
+	/* XXX This is not correct but it is good enough for the
+	 * XXX one place that wants this, IRDA.  If we wanted to
+	 * XXX do this right, we'd need to add a qdisc op to
+	 * XXX probe for the queue state.
+	 */
+	return skb_queue_empty(&txq->qdisc->q);
+}
+
 static inline int __qdisc_enqueue_tail(struct sk_buff *skb, struct Qdisc *sch,
 				       struct sk_buff_head *list)
 {
-- 
1.5.6

Re: [PATCH 13/39]: pkt_sched: Add qdisc_all_tx_empty()

[Patrick McHardy]

David Miller wrote:
> As per comments, this isn't a foolproof implementation.  This is being
> added so that we can contain and isolate all the explicit ->tx_queue
> references in the tree.
> 
> +/* Are all TX queues of the device empty?  */
> +static inline bool qdisc_all_tx_empty(const struct net_device *dev)
> +{
> +	struct netdev_queue *txq = &dev->tx_queue;
> +
> +	/* XXX This is not correct but it is good enough for the
> +	 * XXX one place that wants this, IRDA.  If we wanted to
> +	 * XXX do this right, we'd need to add a qdisc op to
> +	 * XXX probe for the queue state.
> +	 */
> +	return skb_queue_empty(&txq->qdisc->q);
> +}


It this comment referring to the fact that its looking at
qdisc->q itself, while the qdisc might be using internal
queues? If so, just using txq->qdisc->q.qlen should be
fine since qdiscs are required to always update this value,
even if they're not using the queue itself.
--
To unsubscribe from this list: send the line "unsubscribe linux-wireless" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH 13/39]: pkt_sched: Add qdisc_all_tx_empty()

[David Miller]

From: Patrick McHardy <kaber-dcUjhNyLwpNeoWH0uzbU5w@public.gmane.org>
Date: Thu, 03 Jul 2008 15:11:26 +0200

> David Miller wrote:
> > As per comments, this isn't a foolproof implementation.  This is being
> > added so that we can contain and isolate all the explicit ->tx_queue
> > references in the tree.
> > 
> > +/* Are all TX queues of the device empty?  */
> > +static inline bool qdisc_all_tx_empty(const struct net_device *dev)
> > +{
> > +	struct netdev_queue *txq = &dev->tx_queue;
> > +
> > +	/* XXX This is not correct but it is good enough for the
> > +	 * XXX one place that wants this, IRDA.  If we wanted to
> > +	 * XXX do this right, we'd need to add a qdisc op to
> > +	 * XXX probe for the queue state.
> > +	 */
> > +	return skb_queue_empty(&txq->qdisc->q);
> > +}
> 
> 
> It this comment referring to the fact that its looking at
> qdisc->q itself, while the qdisc might be using internal
> queues? If so, just using txq->qdisc->q.qlen should be
> fine since qdiscs are required to always update this value,
> even if they're not using the queue itself.

Indeed, and this is exactly what the sch_generic.c packet output path
uses as a test too.

Thanks Patrick, I'll fix this up.
--
To unsubscribe from this list: send the line "unsubscribe linux-wireless" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html