From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1756640Ab1AaWkW (ORCPT ); Mon, 31 Jan 2011 17:40:22 -0500 Received: from e33.co.us.ibm.com ([32.97.110.151]:45337 "EHLO e33.co.us.ibm.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754069Ab1AaWkU (ORCPT ); Mon, 31 Jan 2011 17:40:20 -0500 Subject: Re: kswapd hung tasks in 2.6.38-rc1 From: Dave Hansen To: CAI Qian Cc: linux-mm , linux-kernel@vger.kernel.org, aarcange In-Reply-To: <1296507528.7797.4609.camel@nimitz> References: <1150342867.83404.1295513748640.JavaMail.root@zmail06.collab.prod.int.phx2.redhat.com> <1296507528.7797.4609.camel@nimitz> Content-Type: text/plain; charset="ANSI_X3.4-1968" Date: Mon, 31 Jan 2011 14:40:16 -0800 Message-ID: <1296513616.7797.4929.camel@nimitz> Mime-Version: 1.0 X-Mailer: Evolution 2.30.3 Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Mon, 2011-01-31 at 12:58 -0800, Dave Hansen wrote: > On Thu, 2011-01-20 at 03:55 -0500, CAI Qian wrote: > > When running LTP oom01 [1] testing, the allocation process stopped > > processing right after starting to swap. > > I'm seeing the same stuff, but on -rc2. I thought it was > transparent-hugepage-related, but I don't see much of a trace of it in > the stack dumps. > > http://sr71.net/~dave/ibm/config-v2.6.38-rc2 > > It happened to me as well around the time that things started to hit > swap. Still not a very good data point, but I ran a heavy swap load for an hour or so without reproducing this. But, it happened again after I enabled transparent huge pages. I managed to get a sysrq-t dump out of it: http://sr71.net/~dave/ibm/2.6.38-rc2-hang-0.txt khugepaged is one of the three running tasks. Note, I set both its sleep timeouts to zero to stress it out a bit. I'll keep trying to reproduce without THP. -- Dave