On Tue, 2007-12-18 at 07:34 -0500, Mark Seger wrote: > plotting with gnuplot. As it turned out, I would very frequently see > spikes of 200MB/sec on my 1Gb link. A colleague noticed the reason was > because the network counters were being updated every 0.9765 seconds and > this was causing the problem. Hello Mark, when you're measuring performance in small time frame, you couldn't see packets for a while due the new NAPI scheme. When a device is using NAPI in a heavy traffic environment, the interrupts are disabled and the packets are got in a polling manner, through dev->poll(). Since dev->poll() is triggered by ksoftirqd, which is a normal lower priority userspace application, and if you have a heavy load machine, the packets could be delayed until ksoftirqd is scheduled again. I am also aware that they backported NAPI to 2.4 kernel, probably 2.4.20, but I am not sure. -- Breno Leitão Linux eServer IO, LTC Brazil leitao@xxxxxxxxxxxxxxxxxx - To unsubscribe from this list: send the line "unsubscribe linux-net" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
