From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S264561AbUGHVlz (ORCPT ); Thu, 8 Jul 2004 17:41:55 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S263761AbUGHVly (ORCPT ); Thu, 8 Jul 2004 17:41:54 -0400 Received: from grendel.digitalservice.pl ([217.67.200.140]:39300 "HELO mail.digitalservice.pl") by vger.kernel.org with SMTP id S264561AbUGHVln (ORCPT ); Thu, 8 Jul 2004 17:41:43 -0400 From: "R. J. Wysocki" Organization: SiSK To: Andrew Morton Subject: Re: 2.6.7-mm6: system hang with KAMix on dual Opteron w/ NUMA Date: Thu, 8 Jul 2004 23:50:52 +0200 User-Agent: KMail/1.5 Cc: linux-kernel@vger.kernel.org References: <200407082324.55181.rjwysocki@sisk.pl> <20040708142153.60c2cb69.akpm@osdl.org> In-Reply-To: <20040708142153.60c2cb69.akpm@osdl.org> MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-2" Content-Transfer-Encoding: 7bit Content-Disposition: inline Message-Id: <200407082350.52483.rjwysocki@sisk.pl> Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org On Thursday 08 of July 2004 23:21, Andrew Morton wrote: > "R. J. Wysocki" wrote: > > I've just discovered that I'm (consistently) able to hang the system > > solid by trying to adjust the sound volume using KAMix on 2.6.7-mm6 (SuSE > > 9.1/ AMD64). > > Please ensure that your kernel has this patch: > > --- 25/sound/core/control.c~snd_ctl_read-locking-fix-fix 2004-07-06 > 21:27:51.618683360 -0700 +++ 25-akpm/sound/core/control.c 2004-07-06 > 21:27:51.622682752 -0700 @@ -1116,7 +1116,7 @@ static ssize_t > snd_ctl_read(struct file > wait_queue_t wait; > if ((file->f_flags & O_NONBLOCK) != 0 || result > 0) { > err = -EAGAIN; > - goto out; > + goto __end; > } > init_waitqueue_entry(&wait, current); > add_wait_queue(&ctl->change_sleep, &wait); > @@ -1137,7 +1137,7 @@ static ssize_t snd_ctl_read(struct file > kfree(kev); > if (copy_to_user(buffer, &ev, sizeof(snd_ctl_event_t))) { > err = -EFAULT; > - goto __end; > + goto out; > } > spin_lock_irq(&ctl->read_lock); > buffer += sizeof(snd_ctl_event_t); > _ It hadn't and the patch fixed the issue. :-) Thanks, rjw -- Rafael J. Wysocki ---------------------------- For a successful technology, reality must take precedence over public relations, for nature cannot be fooled. -- Richard P. Feynman