[PATCH 26/39]: netdev: netdev_priv() can now be sane again.

[PATCH 26/39]: netdev: netdev_priv() can now be sane again.

[David Miller]

The private area of a netdev is now at a fixed offset once more.

Unfortunately, some assumptions that netdev_priv() == netdev->priv
crept back into the tree.  In particular this happened in the
loopback driver.  Make it use netdev->ml_priv.

Signed-off-by: David S. Miller <davem@davemloft.net>
---
 drivers/net/loopback.c    |    8 ++++----
 include/linux/netdevice.h |    4 +++-
 2 files changed, 7 insertions(+), 5 deletions(-)

diff --git a/drivers/net/loopback.c b/drivers/net/loopback.c
index 41b774b..49f6bc0 100644
--- a/drivers/net/loopback.c
+++ b/drivers/net/loopback.c
@@ -153,7 +153,7 @@ static int loopback_xmit(struct sk_buff *skb, struct net_device *dev)
 	dev->last_rx = jiffies;
 
 	/* it's OK to use per_cpu_ptr() because BHs are off */
-	pcpu_lstats = netdev_priv(dev);
+	pcpu_lstats = dev->ml_priv;
 	lb_stats = per_cpu_ptr(pcpu_lstats, smp_processor_id());
 	lb_stats->bytes += skb->len;
 	lb_stats->packets++;
@@ -171,7 +171,7 @@ static struct net_device_stats *get_stats(struct net_device *dev)
 	unsigned long packets = 0;
 	int i;
 
-	pcpu_lstats = netdev_priv(dev);
+	pcpu_lstats = dev->ml_priv;
 	for_each_possible_cpu(i) {
 		const struct pcpu_lstats *lb_stats;
 
@@ -207,13 +207,13 @@ static int loopback_dev_init(struct net_device *dev)
 	if (!lstats)
 		return -ENOMEM;
 
-	dev->priv = lstats;
+	dev->ml_priv = lstats;
 	return 0;
 }
 
 static void loopback_dev_free(struct net_device *dev)
 {
-	struct pcpu_lstats *lstats = netdev_priv(dev);
+	struct pcpu_lstats *lstats = dev->ml_priv;
 
 	free_percpu(lstats);
 	free_netdev(dev);
diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h
index df44e44..8558046 100644
--- a/include/linux/netdevice.h
+++ b/include/linux/netdevice.h
@@ -809,7 +809,9 @@ void dev_net_set(struct net_device *dev, struct net *net)
  */
 static inline void *netdev_priv(const struct net_device *dev)
 {
-	return dev->priv;
+	return (char *)dev + ((sizeof(struct net_device)
+			       + NETDEV_ALIGN_CONST)
+			      & ~NETDEV_ALIGN_CONST);
 }
 
 /* Set the sysfs physical device reference for the network logical device
-- 
1.5.6

Re: [PATCH 26/39]: netdev: netdev_priv() can now be sane again.

[Krzysztof Halasa]

David Miller <davem@davemloft.net> writes:

> The private area of a netdev is now at a fixed offset once more.
>
> Unfortunately, some assumptions that netdev_priv() == netdev->priv
> crept back into the tree.  In particular this happened in the
> loopback driver.

Perhaps the netdev_priv() should be renamed to avoid confusion?
-- 
Krzysztof Halasa

Re: [PATCH 26/39]: netdev: netdev_priv() can now be sane again.

[David Miller]

From: Krzysztof Halasa <khc@pm.waw.pl>
Date: Thu, 03 Jul 2008 13:23:32 +0200

> David Miller <davem@davemloft.net> writes:
> 
> > The private area of a netdev is now at a fixed offset once more.
> >
> > Unfortunately, some assumptions that netdev_priv() == netdev->priv
> > crept back into the tree.  In particular this happened in the
> > loopback driver.
> 
> Perhaps the netdev_priv() should be renamed to avoid confusion?

If all drivers call netdev_priv(), there is no confusion :)

I plan to rename dev->priv to dev->_priv and fix all the fallout.

Re: [PATCH 26/39]: netdev: netdev_priv() can now be sane again.

[Krzysztof Halasa]

David Miller <davem@davemloft.net> writes:

> If all drivers call netdev_priv(), there is no confusion :)
>
> I plan to rename dev->priv to dev->_priv and fix all the fallout.

Do you mean practically deleting dev->priv? That would make sense.
-- 
Krzysztof Halasa

Re: [PATCH 26/39]: netdev: netdev_priv() can now be sane again.

[David Miller]

From: Krzysztof Halasa <khc@pm.waw.pl>
Date: Thu, 03 Jul 2008 16:18:21 +0200

> David Miller <davem@davemloft.net> writes:
> 
> > If all drivers call netdev_priv(), there is no confusion :)
> >
> > I plan to rename dev->priv to dev->_priv and fix all the fallout.
> 
> Do you mean practically deleting dev->priv? That would make sense.

Yes, that's the basic idea.  It's probably the only way to
prevent this problem in the future.