netdev
[Top] [All Lists]

Re: e1000 w/ NAPI + SMP = 99% CPU utilization

To: Robert Olsson <Robert.Olsson@xxxxxxxxxxx>
Subject: Re: e1000 w/ NAPI + SMP = 99% CPU utilization
From: P@xxxxxxxxxxxxxx
Date: Wed, 09 Jun 2004 10:01:02 +0100
Cc: Chris Carpinello <chriscarpinello@xxxxxxxxxxx>, netdev@xxxxxxxxxxx
In-reply-to: <16582.49526.374109.580312@xxxxxxxxxxxx>
References: <BAY1-F139xaq8HCdrWi0004e64b@xxxxxxxxxxx> <16582.49526.374109.580312@xxxxxxxxxxxx>
Sender: netdev-bounce@xxxxxxxxxxx
User-agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.6) Gecko/20040124
Robert Olsson wrote:
Chris Carpinello writes:

Hello!

Is seems like your network load @ ~202 Mbps gets you system into continuing polling as we see very few interrupts on your eth3.
This means that rx_softirq reschedules itself do_softirq() kicks
ksoftird to prevent the rx_softirq from monopolize the system.
So now all the work gets accounted in ksoftird And by design ->poll is strictly serialized per device to guarantee ordering and avoid cache bouncing we only see one ksoftirq used as use only have one input device.

Pádraig suggest binding to separate CPU's. This is normally a good thing but as you only have one input device it will not help.

agreed. All traffic is on eth3 so you can't share it over CPUs

And didn't we just see a fix for ifconfig down oops?

yep, seems like it:
http://marc.theaimsgroup.com/?l=linux-netdev&m=108631346103966&w=2

Pádraig.


<Prev in Thread] Current Thread [Next in Thread>