Received: with ECARTIS (v1.0.0; list netdev); Tue, 27 Jul 2004 13:24:24 -0700 (PDT) Received: from mail1.slu.se (mail1.slu.se [130.238.96.11]) by oss.sgi.com (8.13.0/8.13.0) with ESMTP id i6RKOIkq029215 for ; Tue, 27 Jul 2004 13:24:19 -0700 Received: from robur.slu.se (robur.slu.se [130.238.98.12]) by mail1.slu.se (8.12.10/8.12.10) with ESMTP id i6RKO1IR009741; Tue, 27 Jul 2004 22:24:02 +0200 Received: by robur.slu.se (Postfix, from userid 1000) id CFAD5EC33E; Tue, 27 Jul 2004 22:24:01 +0200 (CEST) From: Robert Olsson MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Message-ID: <16646.47585.814327.628319@robur.slu.se> Date: Tue, 27 Jul 2004 22:24:01 +0200 To: Pasi Sjoholm Cc: Robert Olsson , Francois Romieu , H?ctor Mart?n , Linux-Kernel , , , , Subject: Re: ksoftirqd uses 99% CPU triggered by network traffic (maybe RLT-8139 related) In-Reply-To: References: <16646.14381.740376.204381@robur.slu.se> X-Mailer: VM 7.18 under Emacs 21.3.1 X-archive-position: 7208 X-ecartis-version: Ecartis v1.0.0 Sender: netdev-bounce@oss.sgi.com Errors-to: netdev-bounce@oss.sgi.com X-original-sender: Robert.Olsson@data.slu.se Precedence: bulk X-list: netdev Pasi Sjoholm writes: > Yeah, when the ksoftirqd is taking all the cpu it will be like that, but > when the kernel is behaving normally the starving diff is between 0->1sec. Well ksoftirqd makes your kernel load just visible which is good and ksofirqd gets accounted for this when softirq's get deferred to it. It may look like goes from 0 to 100% but thats probably not the case. The problem is we can starve userland at high loads. As said we were trying some way to cure this I may have some old patch if you like to try. Cheers. --ro