kernel performance update - 2.6.14-rc3
Quoted 6 lines · Kenneth W
Even though
softirq is invoked at the end ofdev_queue_xmit()vialocal_bh_enable(), not all execution of softirq will result a__wake_up(). With higher HZ rate, timer interrupt is more frequent and thus more softirq invocation and leads to more__wake_up(), which then takes us to higher throughput because cpu spend less time in idle.
Since the loopback xmit->rx path probably isn't being called in interrupt context might something like the patch below be needed?
Please forgive me if this is wrong, i've not even tried compiling this change let alone tested it.
Jon
drivers/net/loopback.c renamed from drivers/net/loopback.c-orig +1 -1
@@ -153,7 +153,7 @@ static int loopback_xmit(struct sk_buff lb_stats->tx_packets++;put_cpu (); -netif_rx (skb); +netif_rx_ni (skb);return (0 ); }
Jon Burgess wrote on Friday, October 07, 2005 5:50 PM
Quoted 8 lines
Quoted 6 lines · Kenneth W
Even though
softirq is invoked at the end ofdev_queue_xmit()vialocal_bh_enable(), not all execution of softirq will result a__wake_up(). With higher HZ rate, timer interrupt is more frequent and thus more softirq invocation and leads to more__wake_up(), which then takes us to higher throughput because cpu spend less time in idle.
Quoted 5 lines
Since the loopback
xmit->rxpath probably isn't being called in interrupt context might something like the patch below be needed?Please forgive me if this is wrong, i've not even tried compiling this change let alone tested it.
I don't think this patch has any effect. dev_queue_xmit turns off local_bh when calling loopback_xmit, so calls to do_softirq from netif_rx_ni Will be a noop since do_softirq immediately return upon seeing in_interrupt().
- Ken