From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S264387AbTLPXuu (ORCPT ); Tue, 16 Dec 2003 18:50:50 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S264389AbTLPXuu (ORCPT ); Tue, 16 Dec 2003 18:50:50 -0500 Received: from e35.co.us.ibm.com ([32.97.110.133]:49847 "EHLO e35.co.us.ibm.com") by vger.kernel.org with ESMTP id S264387AbTLPXut (ORCPT ); Tue, 16 Dec 2003 18:50:49 -0500 Subject: Re: Double Interrupt with HT From: john stultz To: Miroslaw KLABA Cc: lkml In-Reply-To: <1071615336.3fdf8d6840208@ssl0.ovh.net> References: <20031215155843.210107b6.totoro@totoro.be> <1071603069.991.194.camel@cog.beaverton.ibm.com> <1071615336.3fdf8d6840208@ssl0.ovh.net> Content-Type: text/plain Message-Id: <1071618630.1013.11.camel@cog.beaverton.ibm.com> Mime-Version: 1.0 X-Mailer: Ximian Evolution 1.4.5 (1.4.5-7) Date: Tue, 16 Dec 2003 15:50:31 -0800 Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org On Tue, 2003-12-16 at 14:55, Miroslaw KLABA wrote: > I had the problem with 2.4.22, 2.4.22-ac4, 2.4.23 and 2.4.24-pre1. Ok, so its been around awhile. Do you remember what was the last 2.4 kernel where you did not see this problem? > The problem is that all the kernel is working "twice the speed". > The command "while true; do date; sleep 1; done;" shows that the date is growing > 2 seconds per second... :/ > I found a patch for irqbalance for 2.4.23, and now I don't have the problem > anymore with the clock. > http://www.hardrock.org/kernel/2.4.23/irqbalance-2.4.23-jb.patch Hmm. Just skimming that patch, I notice it won't work on clustered apic systems. They've dropped the following chunk from set_ioapic_affinity and forgot to re-add it. - /* pick a single cpu for clustered xapics */ - if(clustered_apic_mode == CLUSTERED_APIC_XAPIC){ - int cpu = ffs(mask)-1; - mask = cpu_to_physical_apicid(cpu); - } Further I can't see how it fixes the problem, but it may just be working around the issue. I'd be interested in what the patch author thinks. > I think it is a bug with the via chipset, but I'm not able to get deeper in the > kernel code. Could be, but I suspect interrupt routing isn't happening properly at boot time. The irqbalance code just forces it to be readjusted correctly once your up and running. thanks -john