Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S268849AbUIHEWm (ORCPT ); Wed, 8 Sep 2004 00:22:42 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S268889AbUIHEWm (ORCPT ); Wed, 8 Sep 2004 00:22:42 -0400 Received: from www2.muking.org ([216.231.42.228]:16413 "HELO www2.muking.org") by vger.kernel.org with SMTP id S268836AbUIHEWN (ORCPT ); Wed, 8 Sep 2004 00:22:13 -0400 To: linux-kernel@vger.kernel.org Subject: Re: [patch] voluntary-preempt-2.6.9-rc1-bk12-R6 From: Kevin Hilman Organization: None to speak of. Date: 07 Sep 2004 21:22:11 -0700 Message-ID: <834qm92xvw.fsf@www2.muking.org> User-Agent: Gnus/5.09 (Gnus v5.9.0) Emacs/21.3 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 974 Lines: 19 I'm running the VP patch on a PII 400MHz to closer approximate an embedded target. I get a 21ms latency trace during boot which dwarfs other latencies and prevents me from seeing any of the later latencies when I'm running my test. The trace (from -R5) is available here: http://hilman.org/kevin/VP/trace-cond_resched.txt At first glance, it appears to be the result of an accumulation of calls to __delay() from the 3c59x vortex driver. Any ideas what's going on here? Is there a way to disable the trace by default and enable it later via /proc? I see that the preemption itself can be disabled via command-line and then enable later via /proc but I don't see the same for the latency trace. Kevin http://hilman.org/ - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/