2010-04-02 04:10:40

by FUJITA Tomonori

Subject: [PATCH] mwl8k: use the dma state API instead of the pci equivalents

The DMA API is preferred.

No functional change.

Signed-off-by: FUJITA Tomonori <[email protected]>
---
drivers/net/wireless/mwl8k.c | 12 ++++++------
1 files changed, 6 insertions(+), 6 deletions(-)

diff --git a/drivers/net/wireless/mwl8k.c b/drivers/net/wireless/mwl8k.c
index ac65e13..fe3c0e0 100644
--- a/drivers/net/wireless/mwl8k.c
+++ b/drivers/net/wireless/mwl8k.c
@@ -108,7 +108,7 @@ struct mwl8k_rx_queue {
dma_addr_t rxd_dma;
struct {
struct sk_buff *skb;
- DECLARE_PCI_UNMAP_ADDR(dma)
+ DEFINE_DMA_UNMAP_ADDR(dma);
} *buf;
};

@@ -964,7 +964,7 @@ static int rxq_refill(struct ieee80211_hw *hw, int index, int limit)
if (rxq->tail == MWL8K_RX_DESCS)
rxq->tail = 0;
rxq->buf[rx].skb = skb;
- pci_unmap_addr_set(&rxq->buf[rx], dma, addr);
+ dma_unmap_addr_set(&rxq->buf[rx], dma, addr);

rxd = rxq->rxd + (rx * priv->rxd_ops->rxd_size);
priv->rxd_ops->rxd_refill(rxd, addr, MWL8K_RX_MAXSZ);
@@ -985,9 +985,9 @@ static void mwl8k_rxq_deinit(struct ieee80211_hw *hw, int index)
for (i = 0; i < MWL8K_RX_DESCS; i++) {
if (rxq->buf[i].skb != NULL) {
pci_unmap_single(priv->pdev,
- pci_unmap_addr(&rxq->buf[i], dma),
+ dma_unmap_addr(&rxq->buf[i], dma),
MWL8K_RX_MAXSZ, PCI_DMA_FROMDEVICE);
- pci_unmap_addr_set(&rxq->buf[i], dma, 0);
+ dma_unmap_addr_set(&rxq->buf[i], dma, 0);

kfree_skb(rxq->buf[i].skb);
rxq->buf[i].skb = NULL;
@@ -1061,9 +1061,9 @@ static int rxq_process(struct ieee80211_hw *hw, int index, int limit)
rxq->buf[rxq->head].skb = NULL;

pci_unmap_single(priv->pdev,
- pci_unmap_addr(&rxq->buf[rxq->head], dma),
+ dma_unmap_addr(&rxq->buf[rxq->head], dma),
MWL8K_RX_MAXSZ, PCI_DMA_FROMDEVICE);
- pci_unmap_addr_set(&rxq->buf[rxq->head], dma, 0);
+ dma_unmap_addr_set(&rxq->buf[rxq->head], dma, 0);

rxq->head++;
if (rxq->head == MWL8K_RX_DESCS)
--
1.7.0

2010-04-06 21:15:13

by John W. Linville

Subject: Re: [PATCH] mwl8k: use the dma state API instead of the pci equivalents

On Fri, Apr 02, 2010 at 01:10:38PM +0900, FUJITA Tomonori wrote:
> The DMA API is preferred.
>
> No functional change.
>
> Signed-off-by: FUJITA Tomonori <[email protected]>

Again, what tree is this against? mwl8k fails to build after this patch.

John
--
John W. Linville Someday the world will need a hero, and you
[email protected] might be all we have. Be ready.

2010-04-02 04:52:29

by Lennert Buytenhek

Subject: Re: [PATCH] mwl8k: use the dma state API instead of the pci equivalents

On Fri, Apr 02, 2010 at 01:10:38PM +0900, FUJITA Tomonori wrote:

> The DMA API is preferred.
>
> No functional change.
>
> Signed-off-by: FUJITA Tomonori <[email protected]>

Acked-by: Lennert Buytenhek <[email protected]>

2010-04-07 00:35:30

by FUJITA Tomonori

Subject: Re: [PATCH] mwl8k: use the dma state API instead of the pci equivalents

On Tue, 6 Apr 2010 17:02:28 -0400
"John W. Linville" <[email protected]> wrote:

> On Fri, Apr 02, 2010 at 01:10:38PM +0900, FUJITA Tomonori wrote:
> > The DMA API is preferred.
> >
> > No functional change.
> >
> > Signed-off-by: FUJITA Tomonori <[email protected]>
>
> Again, what tree is this against? mwl8k fails to build after this patch.

Oops, sorry. It was against net-next. I've just got wireless-next and
found that it doesn't have the DMA state API yet, which was introduced
on 2.6.34-rc2. When do you plan to update your tree? I'll work on both
patches again later on.

Thanks,