pcnet32: Fix PCnet32 performance bug on non-coherent architecutres

The PCnet32 driver always passed the the size of the largest possible packet
to the pci_dma_sync_single_for_cpu and pci_dma_sync_single_for_device.
This results in a fairly large "colateral damage" in the caches and makes
the flush operation itself much slower.  On a system with a 40MHz CPU this
patch increases network bandwidth by about 12%.
Signed-off-by: NRalf Baechle <ralf@linux-mips.org>
Acked-by: NDon Fry <pcnet32@verizon.net>
Signed-off-by: NJeff Garzik <jeff@garzik.org>

drivers/net/pcnet32.c

@@ -1234,14 +1234,14 @@ static void pcnet32_rx_entry(struct net_device *dev,
 		skb_put(skb, pkt_len);	/* Make room */
 		pci_dma_sync_single_for_cpu(lp->pci_dev,
 					    lp->rx_dma_addr[entry],
-					    PKT_BUF_SZ - 2,
+					    pkt_len,
 					    PCI_DMA_FROMDEVICE);
 		eth_copy_and_sum(skb,
 				 (unsigned char *)(lp->rx_skbuff[entry]->data),
 				 pkt_len, 0);
 		pci_dma_sync_single_for_device(lp->pci_dev,
 					       lp->rx_dma_addr[entry],
-					       PKT_BUF_SZ - 2,
+					       pkt_len,
 					       PCI_DMA_FROMDEVICE);
 	}
 	lp->stats.rx_bytes += skb->len;