Donate to e Foundation | Murena handsets with /e/OS | Own a part of Murena! Learn more

Commit b2cbbd8e authored by Don Fry's avatar Don Fry Committed by Jeff Garzik
Browse files

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: default avatarRalf Baechle <ralf@linux-mips.org>
Acked-by: default avatarDon Fry <pcnet32@verizon.net>
Signed-off-by: default avatarJeff Garzik <jeff@garzik.org>
parent 4c44fd00

drivers/net/pcnet32.c

+2 −2
Original line number Diff line number Diff line
@@ -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;