From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S269139AbUIHMJH (ORCPT ); Wed, 8 Sep 2004 08:09:07 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S269152AbUIHMJG (ORCPT ); Wed, 8 Sep 2004 08:09:06 -0400 Received: from cantor.suse.de ([195.135.220.2]:42119 "EHLO Cantor.suse.de") by vger.kernel.org with ESMTP id S269139AbUIHMEp (ORCPT ); Wed, 8 Sep 2004 08:04:45 -0400 To: Pavel Machek Cc: linux-kernel@vger.kernel.org Subject: Re: Fw: 2.6.9-rc1-mm4: swsusp + AMD64 = LOCKUP on CPU0 References: <20040908021637.57525d43.akpm@osdl.org.suse.lists.linux.kernel> <20040908102652.GA2921@atrey.karlin.mff.cuni.cz.suse.lists.linux.kernel> From: Andi Kleen Date: 08 Sep 2004 14:01:08 +0200 In-Reply-To: <20040908102652.GA2921@atrey.karlin.mff.cuni.cz.suse.lists.linux.kernel> Message-ID: User-Agent: Gnus/5.09 (Gnus v5.9.0) Emacs/21.2 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 Pavel Machek writes: > Hi! > > > One for you guys on lkml ;) > > It simply takes long to count pages (O(n^2) algorithm), so watchdog > triggers. I have better algorithm locally, but would like merge to > linus first. (I posted it to lkml some days ago, I can attach the > bigdiff). > > Just disable the watchdog. Suspend *is* going to take time with > disabled interrupts. As a short term workaround you could also add touch_nmi_watchdog()s in that loop. -Andi