diff options
author | Eric Dumazet <eric.dumazet@gmail.com> | 2009-05-28 00:00:41 (GMT) |
---|---|---|
committer | David S. Miller <davem@davemloft.net> | 2009-05-29 08:46:27 (GMT) |
commit | cdd0db058d64f2619b08e7d098ee0014aa0a5bc1 (patch) | |
tree | a5e31f6687eff581309b095781c7e2f0bc676182 /drivers/net/yellowfin.c | |
parent | 28679751a924c11f7135641f26e99249385de5b4 (diff) | |
download | linux-cdd0db058d64f2619b08e7d098ee0014aa0a5bc1.tar.xz |
net: dont update dev->trans_start
Second round of drivers for Gb cards (and NIU one I forgot in the 10GB round)
Now that core network takes care of trans_start updates, dont do it
in drivers themselves, if possible. Drivers can avoid one cache miss
(on dev->trans_start) in their start_xmit() handler.
Exceptions are NETIF_F_LLTX drivers
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Diffstat (limited to 'drivers/net/yellowfin.c')
-rw-r--r-- | drivers/net/yellowfin.c | 3 |
1 files changed, 1 insertions, 2 deletions
diff --git a/drivers/net/yellowfin.c b/drivers/net/yellowfin.c index 7477ffd..3c7a505 100644 --- a/drivers/net/yellowfin.c +++ b/drivers/net/yellowfin.c @@ -717,7 +717,7 @@ static void yellowfin_tx_timeout(struct net_device *dev) if (yp->cur_tx - yp->dirty_tx < TX_QUEUE_SIZE) netif_wake_queue (dev); /* Typical path */ - dev->trans_start = jiffies; + dev->trans_start = jiffies; /* prevent tx timeout */ dev->stats.tx_errors++; } @@ -876,7 +876,6 @@ static int yellowfin_start_xmit(struct sk_buff *skb, struct net_device *dev) netif_start_queue (dev); /* Typical path */ else yp->tx_full = 1; - dev->trans_start = jiffies; if (yellowfin_debug > 4) { printk(KERN_DEBUG "%s: Yellowfin transmit frame #%d queued in slot %d.\n", |