[PATCH] vxge: use pci_dma_mapping_error to test return value

[PATCH] vxge: use pci_dma_mapping_error to test return value

[Denis Kirjanov <kirjanov@gmail.com]

pci_dma_mapping_error should be used to test return value of
pci_map_single or pci_map_page.

Signed-off-by: Denis Kirjanov <kirjanov@gmail.com>
---

drivers/net/vxge/vxge-main.c |    4 ++--
 1 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/net/vxge/vxge-main.c b/drivers/net/vxge/vxge-main.c
index 0fdfd58..f7ed495 100644
--- a/drivers/net/vxge/vxge-main.c
+++ b/drivers/net/vxge/vxge-main.c
@@ -310,7 +310,7 @@ static int vxge_rx_map(void *dtrh, struct vxge_ring *ring)
 	dma_addr = pci_map_single(ring->pdev, rx_priv->skb_data,
 				rx_priv->data_size, PCI_DMA_FROMDEVICE);
 
-	if (dma_addr == 0) {
+	if (unlikely(pci_dma_mapping_error(ring->pdev, dma_addr))) {
 		ring->stats.pci_map_fail++;
 		return -EIO;
 	}
@@ -821,7 +821,7 @@ vxge_xmit(struct sk_buff *skb, struct net_device *dev)
 	int frg_cnt, first_frg_len;
 	skb_frag_t *frag;
 	int i = 0, j = 0, avail;
-	u64 dma_pointer;
+	dma_addr_t dma_pointer;
 	struct vxge_tx_priv *txdl_priv = NULL;
 	struct __vxge_hw_fifo *fifo_hw;
 	int offload_type;

Re: [PATCH] vxge: use pci_dma_mapping_error to test return value

[David Miller]

From: "Denis Kirjanov <kirjanov@gmail.com" <kirjanov@gmail.com>
Date: Wed, 6 Jan 2010 00:16:59 +0300

> @@ -821,7 +821,7 @@ vxge_xmit(struct sk_buff *skb, struct net_device *dev)
>  	int frg_cnt, first_frg_len;
>  	skb_frag_t *frag;
>  	int i = 0, j = 0, avail;
> -	u64 dma_pointer;
> +	dma_addr_t dma_pointer;
>  	struct vxge_tx_priv *txdl_priv = NULL;
>  	struct __vxge_hw_fifo *fifo_hw;
>  	int offload_type;

This will cause build warnings on platforms withe dma_addr_t
is some type other than "unsigned long long", such as on
sparc64 where it is a u32.

This is because of the %ll printf format used to display it
a few lines down from here.

Re: [PATCH] vxge: use pci_dma_mapping_error to test return value

[Denis Kirjanov <kirjanov@gmail.com]

On Fri, Jan 08, 2010 at 00:55 -0800, David Miller wrote:
> From: "Denis Kirjanov <kirjanov@gmail.com" <kirjanov@gmail.com>
> Date: Wed, 6 Jan 2010 00:16:59 +0300
> 
> > @@ -821,7 +821,7 @@ vxge_xmit(struct sk_buff *skb, struct net_device *dev)
> >  	int frg_cnt, first_frg_len;
> >  	skb_frag_t *frag;
> >  	int i = 0, j = 0, avail;
> > -	u64 dma_pointer;
> > +	dma_addr_t dma_pointer;
> >  	struct vxge_tx_priv *txdl_priv = NULL;
> >  	struct __vxge_hw_fifo *fifo_hw;
> >  	int offload_type;
> 
> This will cause build warnings on platforms withe dma_addr_t
> is some type other than "unsigned long long", such as on
> sparc64 where it is a u32.
> 
> This is because of the %ll printf format used to display it
> a few lines down from here.
My mistake. Fixed one.

pci_dma_mapping_error should be used to test return value of
pci_map_single or pci_map_page.

Signed-off-by: Denis Kirjanov <kirjanov@gmail.com>
---

 drivers/net/vxge/vxge-main.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

diff --git a/drivers/net/vxge/vxge-main.c b/drivers/net/vxge/vxge-main.c
index 0fdfd58..b9685e8 100644
--- a/drivers/net/vxge/vxge-main.c
+++ b/drivers/net/vxge/vxge-main.c
@@ -310,7 +310,7 @@ static int vxge_rx_map(void *dtrh, struct vxge_ring *ring)
 	dma_addr = pci_map_single(ring->pdev, rx_priv->skb_data,
 				rx_priv->data_size, PCI_DMA_FROMDEVICE);
 
-	if (dma_addr == 0) {
+	if (unlikely(pci_dma_mapping_error(ring->pdev, dma_addr))) {
 		ring->stats.pci_map_fail++;
 		return -EIO;
 	}

Re: [PATCH] vxge: use pci_dma_mapping_error to test return value

[David Miller]

From: "Denis Kirjanov <kirjanov@gmail.com" <kirjanov@gmail.com>
Date: Sat, 9 Jan 2010 22:52:22 +0300

> My mistake. Fixed one.
> 
> pci_dma_mapping_error should be used to test return value of
> pci_map_single or pci_map_page.
> 
> Signed-off-by: Denis Kirjanov <kirjanov@gmail.com>

Applied, thanks.