From d11c13e752c4e34777d33579ee0378e0178ef52d Mon Sep 17 00:00:00 2001 From: "shemminger@osdl.org" Date: Tue, 27 Sep 2005 15:02:56 -0700 Subject: [PATCH] sky2: cleanup interrupt processing The receive processing can be cleaned up by not using local variables to store checksum status, instead just put it directly onto the expected skb. Signed-off-by: Stephen Hemminger Signed-off-by: Jeff Garzik --- drivers/net/sky2.c | 65 +++++++++++++++++++++++------------------------------- 1 file changed, 28 insertions(+), 37 deletions(-) (limited to 'drivers/net/sky2.c') diff --git a/drivers/net/sky2.c b/drivers/net/sky2.c index 0cc57293faf..8527a49e3ff 100644 --- a/drivers/net/sky2.c +++ b/drivers/net/sky2.c @@ -1040,13 +1040,14 @@ out_unlock: * NB: the hardware will tell us about partial completion of multi-part * buffers; these are defered until completion. */ -static void sky2_tx_complete(struct net_device *dev, u16 done) +static void sky2_tx_complete(struct sky2_port *sky2, u16 done) { - struct sky2_port *sky2 = netdev_priv(dev); + struct net_device *dev = sky2->netdev; unsigned i; - if (netif_msg_tx_done(sky2)) - printk(KERN_DEBUG "%s: tx done, upto %u\n", dev->name, done); + if (unlikely(netif_msg_tx_done(sky2))) + printk(KERN_DEBUG "%s: tx done, upto %u\n", + dev->name, done); spin_lock(&sky2->tx_lock); @@ -1086,7 +1087,7 @@ out: /* Cleanup all untransmitted buffers, assume transmitter not running */ static inline void sky2_tx_clean(struct sky2_port *sky2) { - sky2_tx_complete(sky2->netdev, sky2->tx_prod); + sky2_tx_complete(sky2, sky2->tx_prod); } /* Network shutdown */ @@ -1421,18 +1422,17 @@ static int sky2_change_mtu(struct net_device *dev, int new_mtu) * For small packets or errors, just reuse existing skb. * For larger pakects, get new buffer. */ -static struct sk_buff *sky2_receive(struct sky2_hw *hw, unsigned port, +static struct sk_buff *sky2_receive(struct sky2_port *sky2, u16 length, u32 status) { - struct net_device *dev = hw->dev[port]; - struct sky2_port *sky2 = netdev_priv(dev); struct ring_info *re = sky2->rx_ring + sky2->rx_next; struct sk_buff *skb = NULL; + struct net_device *dev; const unsigned int bufsize = rx_size(sky2); if (unlikely(netif_msg_rx_status(sky2))) printk(KERN_DEBUG PFX "%s: rx slot %u status 0x%x len %d\n", - dev->name, sky2->rx_next, status, length); + sky2->netdev->name, sky2->rx_next, status, length); sky2->rx_next = (sky2->rx_next + 1) % sky2->rx_pending; @@ -1451,6 +1451,8 @@ static struct sk_buff *sky2_receive(struct sky2_hw *hw, unsigned port, pci_dma_sync_single_for_cpu(sky2->hw->pdev, re->mapaddr, length, PCI_DMA_FROMDEVICE); memcpy(skb->data, re->skb->data, length); + skb->ip_summed = re->skb->ip_summed; + skb->csum = re->skb->csum; pci_dma_sync_single_for_device(sky2->hw->pdev, re->mapaddr, length, PCI_DMA_FROMDEVICE); } else { @@ -1471,12 +1473,14 @@ static struct sk_buff *sky2_receive(struct sky2_hw *hw, unsigned port, re->maplen = bufsize; } - skb->dev = dev; skb_put(skb, length); + dev = sky2->netdev; + skb->dev = dev; skb->protocol = eth_type_trans(skb, dev); dev->last_rx = jiffies; resubmit: + re->skb->ip_summed = CHECKSUM_NONE; sky2_rx_add(sky2, re); return skb; @@ -1517,59 +1521,46 @@ static inline u16 tx_index(u8 port, u32 status, u16 len) * Both ports share the same status interrupt, therefore there is only * one poll routine. */ -static int sky2_poll(struct net_device *dev, int *budget) +static int sky2_poll(struct net_device *dev0, int *budget) { - struct sky2_port *sky2 = netdev_priv(dev); - struct sky2_hw *hw = sky2->hw; - unsigned int to_do = min(dev->quota, *budget); + struct sky2_hw *hw = ((struct sky2_port *) netdev_priv(dev0))->hw; + unsigned int to_do = min(dev0->quota, *budget); unsigned int work_done = 0; u16 hwidx; - unsigned char summed[2] = { CHECKSUM_NONE, CHECKSUM_NONE }; - unsigned int csum[2]; hwidx = sky2_read16(hw, STAT_PUT_IDX); BUG_ON(hwidx >= STATUS_RING_SIZE); rmb(); while (hw->st_idx != hwidx && work_done < to_do) { struct sky2_status_le *le = hw->st_le + hw->st_idx; + struct sky2_port *sky2; struct sk_buff *skb; - u8 port; u32 status; u16 length; + BUG_ON(le->link >= hw->ports); + sky2 = netdev_priv(hw->dev[le->link]); status = le32_to_cpu(le->status); length = le16_to_cpu(le->length); - port = le->link; - - BUG_ON(port >= hw->ports || hw->dev[port] == NULL); switch (le->opcode & ~HW_OWNER) { case OP_RXSTAT: - skb = sky2_receive(hw, port, length, status); + skb = sky2_receive(sky2, length, status); if (likely(skb)) { - /* Add hw checksum if available */ - skb->ip_summed = summed[port]; - skb->csum = csum[port]; - netif_receive_skb(skb); ++work_done; } - - /* Clear for next packet */ - csum[port] = 0; - summed[port] = CHECKSUM_NONE; - break; case OP_RXCHKS: - /* Save computed checksum for next rx */ - csum[port] = le16_to_cpu(status & 0xffff); - summed[port] = CHECKSUM_HW; + skb = sky2->rx_ring[sky2->rx_next].skb; + skb->ip_summed = CHECKSUM_HW; + skb->csum = le16_to_cpu(status); break; case OP_TXINDEXLE: - sky2_tx_complete(hw->dev[port], - tx_index(port, status, length)); + sky2_tx_complete(sky2, + tx_index(sky2->port, status, length)); break; case OP_RXTIMESTAMP: @@ -1599,7 +1590,7 @@ static int sky2_poll(struct net_device *dev, int *budget) rx_set_put(hw->dev[1]); *budget -= work_done; - dev->quota -= work_done; + dev0->quota -= work_done; if (work_done < to_do) { /* * Another chip workaround, need to restart TX timer if status @@ -1613,7 +1604,7 @@ static int sky2_poll(struct net_device *dev, int *budget) hw->intr_mask |= Y2_IS_STAT_BMU; sky2_write32(hw, B0_IMSK, hw->intr_mask); sky2_read32(hw, B0_IMSK); - netif_rx_complete(dev); + netif_rx_complete(dev0); } return work_done >= to_do; -- cgit v1.2.3