[PATCH] korina: adjust headroom for new skb's also

[PATCH] korina: adjust headroom for new skb's also

[Phil Sutter]

This is copy and paste from the original driver. As skb_reserve() is
also called within korina_alloc_ring() when initially allocating the
receive descriptors, the same should be done when allocating new space
after passing an skb to upper layers.

Signed-off-by: Phil Sutter <n0-1@freewrt.org>
---
 drivers/net/korina.c |    3 +++
 1 files changed, 3 insertions(+), 0 deletions(-)

diff --git a/drivers/net/korina.c b/drivers/net/korina.c
index 67fbdf4..60ae7bf 100644
--- a/drivers/net/korina.c
+++ b/drivers/net/korina.c
@@ -416,6 +416,9 @@ static int korina_rx(struct net_device *dev, int limit)
 			if (devcs & ETH_RX_MP)
 				dev->stats.multicast++;
 
+			/* 16 bit align */
+			skb_reserve(skb_new, 2);
+
 			lp->rx_skb[lp->rx_next_done] = skb_new;
 		}
 
-- 
1.5.6.4

[PATCH] korina: drop leftover assignment

[Phil Sutter]

As the assigned value is being overwritten shortly after, it can be
dropped and so the whole variable definition moved to the start of the
function.

Signed-off-by: Phil Sutter <n0-1@freewrt.org>
---
 drivers/net/korina.c |    3 +--
 1 files changed, 1 insertions(+), 2 deletions(-)

diff --git a/drivers/net/korina.c b/drivers/net/korina.c
index 60ae7bf..75010ca 100644
--- a/drivers/net/korina.c
+++ b/drivers/net/korina.c
@@ -743,6 +743,7 @@ static struct ethtool_ops netdev_ethtool_ops = {
 static void korina_alloc_ring(struct net_device *dev)
 {
 	struct korina_private *lp = netdev_priv(dev);
+	struct sk_buff *skb;
 	int i;
 
 	/* Initialize the transmit descriptors */
@@ -758,8 +759,6 @@ static void korina_alloc_ring(struct net_device *dev)
 
 	/* Initialize the receive descriptors */
 	for (i = 0; i < KORINA_NUM_RDS; i++) {
-		struct sk_buff *skb = lp->rx_skb[i];
-
 		skb = dev_alloc_skb(KORINA_RBSIZE + 2);
 		if (!skb)
 			break;
-- 
1.5.6.4

[PATCH] korina: set dev field of allocated socketbufs

[Phil Sutter]

I'm not sure if this is necessary, but the original driver did it and
apparently also many other drivers do it, so it should not be completely
wrong to do it.

Signed-off-by: Phil Sutter <n0-1@freewrt.org>
---
 drivers/net/korina.c |    2 ++
 1 files changed, 2 insertions(+), 0 deletions(-)

diff --git a/drivers/net/korina.c b/drivers/net/korina.c
index 75010ca..56b4db2 100644
--- a/drivers/net/korina.c
+++ b/drivers/net/korina.c
@@ -419,6 +419,7 @@ static int korina_rx(struct net_device *dev, int limit)
 			/* 16 bit align */
 			skb_reserve(skb_new, 2);
 
+			skb_new->dev = dev;
 			lp->rx_skb[lp->rx_next_done] = skb_new;
 		}
 
@@ -763,6 +764,7 @@ static void korina_alloc_ring(struct net_device *dev)
 		if (!skb)
 			break;
 		skb_reserve(skb, 2);
+		skb->dev = dev;
 		lp->rx_skb[i] = skb;
 		lp->rd_ring[i].control = DMA_DESC_IOD |
 				DMA_COUNT(KORINA_RBSIZE);
-- 
1.5.6.4

Re: [PATCH] korina: drop leftover assignment

[Florian Fainelli]

Le Thursday 15 January 2009 23:29:57 Phil Sutter, vous avez écrit :
> As the assigned value is being overwritten shortly after, it can be
> dropped and so the whole variable definition moved to the start of the
> function.
>
> Signed-off-by: Phil Sutter <n0-1@freewrt.org>

Acked-by: Florian Fainelli <florian@openwrt.org>
-- 
Best regards, Florian Fainelli
Email : florian@openwrt.org
http://openwrt.org
-------------------------------

Re: [PATCH] korina: adjust headroom for new skb's also

[Florian Fainelli]

Le Thursday 15 January 2009 23:29:56 Phil Sutter, vous avez écrit :
> This is copy and paste from the original driver. As skb_reserve() is
> also called within korina_alloc_ring() when initially allocating the
> receive descriptors, the same should be done when allocating new space
> after passing an skb to upper layers.
>
> Signed-off-by: Phil Sutter <n0-1@freewrt.org>

Acked-by: Florian Fainelli <florian@openwrt.org>
-- 
Best regards, Florian Fainelli
Email : florian@openwrt.org
http://openwrt.org
-------------------------------

Re: [PATCH] korina: set dev field of allocated socketbufs

[Florian Fainelli]

Le Thursday 15 January 2009 23:29:58 Phil Sutter, vous avez écrit :
> I'm not sure if this is necessary, but the original driver did it and
> apparently also many other drivers do it, so it should not be completely
> wrong to do it.
>
> Signed-off-by: Phil Sutter <n0-1@freewrt.org>

Acked-by: Florian Fainelli <florian@openwrt.org>
-- 
Best regards, Florian Fainelli
Email : florian@openwrt.org
http://openwrt.org
-------------------------------

Re: [PATCH] korina: drop leftover assignment

[David Miller]

From: Florian Fainelli <florian@openwrt.org>
Date: Mon, 19 Jan 2009 22:07:01 +0100

> Le Thursday 15 January 2009 23:29:57 Phil Sutter, vous avez écrit :
> > As the assigned value is being overwritten shortly after, it can be
> > dropped and so the whole variable definition moved to the start of the
> > function.
> >
> > Signed-off-by: Phil Sutter <n0-1@freewrt.org>
> 
> Acked-by: Florian Fainelli <florian@openwrt.org>

Applied.

Re: [PATCH] korina: adjust headroom for new skb's also

[David Miller]

From: Florian Fainelli <florian@openwrt.org>
Date: Mon, 19 Jan 2009 22:07:45 +0100

> Le Thursday 15 January 2009 23:29:56 Phil Sutter, vous avez écrit :
> > This is copy and paste from the original driver. As skb_reserve() is
> > also called within korina_alloc_ring() when initially allocating the
> > receive descriptors, the same should be done when allocating new space
> > after passing an skb to upper layers.
> >
> > Signed-off-by: Phil Sutter <n0-1@freewrt.org>
> 
> Acked-by: Florian Fainelli <florian@openwrt.org>

Applied.

Re: [PATCH] korina: set dev field of allocated socketbufs

[David Miller]

From: Florian Fainelli <florian@openwrt.org>
Date: Mon, 19 Jan 2009 22:08:00 +0100

> Le Thursday 15 January 2009 23:29:58 Phil Sutter, vous avez écrit :
> > I'm not sure if this is necessary, but the original driver did it and
> > apparently also many other drivers do it, so it should not be completely
> > wrong to do it.
> >
> > Signed-off-by: Phil Sutter <n0-1@freewrt.org>
> 
> Acked-by: Florian Fainelli <florian@openwrt.org>

The eth_type_trans() call does the 'skb->dev = dev;' assignment
for you.

Not applied.