From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S263040AbUGXX1x (ORCPT ); Sat, 24 Jul 2004 19:27:53 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S263093AbUGXX1x (ORCPT ); Sat, 24 Jul 2004 19:27:53 -0400 Received: from electric-eye.fr.zoreil.com ([213.41.134.224]:63457 "EHLO fr.zoreil.com") by vger.kernel.org with ESMTP id S263040AbUGXX1w (ORCPT ); Sat, 24 Jul 2004 19:27:52 -0400 Date: Sun, 25 Jul 2004 01:27:12 +0200 From: Francois Romieu To: H?ctor Mart?n Cc: Linux-Kernel Subject: Re: ksoftirqd uses 99% CPU triggered by network traffic Message-ID: <20040725012712.A15785@electric-eye.fr.zoreil.com> References: <4102CF17.2010207@marcansoft.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.2.5.1i In-Reply-To: <4102CF17.2010207@marcansoft.com>; from hector@marcansoft.com on Sat, Jul 24, 2004 at 11:05:27PM +0200 X-Organisation: Land of Sunshine Inc. Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org H?ctor Mart?n : [...] > Interrupt 9 is surely busy, no USB hardware plugged in just in case you're > wondering, and normally (while not using eth2) interrupt 9 is rock solid > (i.e. I doubt ACPI interrupts at all during normal use unless e.g. the power > button is pressed.) At 60 kirq/s without any network traffic, you may disable acpi then usb and eventually poke your nose in the bios setup first. No joke. -- Ueimor