From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S964975AbWHXPSx (ORCPT ); Thu, 24 Aug 2006 11:18:53 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S964986AbWHXPSx (ORCPT ); Thu, 24 Aug 2006 11:18:53 -0400 Received: from ns2.suse.de ([195.135.220.15]:20625 "EHLO mx2.suse.de") by vger.kernel.org with ESMTP id S964975AbWHXPSx (ORCPT ); Thu, 24 Aug 2006 11:18:53 -0400 From: Andi Kleen To: schwidefsky@de.ibm.com Subject: Re: [patch] dubious process system time. Date: Thu, 24 Aug 2006 17:18:29 +0200 User-Agent: KMail/1.9.3 Cc: linux-kernel@vger.kernel.org References: <20060824121825.GA4425@skybase> <1156426103.28464.29.camel@localhost> In-Reply-To: <1156426103.28464.29.camel@localhost> MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-15" Content-Transfer-Encoding: 7bit Content-Disposition: inline Message-Id: <200608241718.29406.ak@suse.de> Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org > At the moment hardirq+softirq is just added to a random process, in > general this is completely wrong. It's better than not accounting it at all. > You just need a system with a cpu hog > and an i/o bound process and you get queer results. Yes, but system load that is invisible to standard monitoring tools is even worse. If you stop accounting it to random processes you have to account it somewhere else. Preferably somewhere that standard tools automatically pick up. -Andi