From mboxrd@z Thu Jan 1 00:00:00 1970 From: Tejun Heo Subject: Re: Long delays while hibernating Date: Wed, 23 Dec 2009 16:47:05 +0900 Message-ID: <4B31CAF9.8070104@kernel.org> References: <74fd948d0912221950m1ade8475pfe3ec3018e14d834@mail.gmail.com> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Return-path: Received: from hera.kernel.org ([140.211.167.34]:49395 "EHLO hera.kernel.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751502AbZLWHnk (ORCPT ); Wed, 23 Dec 2009 02:43:40 -0500 In-Reply-To: <74fd948d0912221950m1ade8475pfe3ec3018e14d834@mail.gmail.com> Sender: linux-ide-owner@vger.kernel.org List-Id: linux-ide@vger.kernel.org To: Pedro Ribeiro Cc: linux-ide@vger.kernel.org On 12/23/2009 12:50 PM, Pedro Ribeiro wrote: > Hi, > > I have a 2.6.32.2 kernel with the TuxOnIce and KDB patches. When I > hibernate, the penultimate phase is the "Atomic copy/restore". On > normal conditions, this takes 1 second or less. > > I have a Thinkpad T400 with an ultrabay, which is a hot-swappable > drive enclosure, which can take in an optical drive (my case), hard > disk or extra battery. > To save some battery, I use the script here > http://www.thinkwiki.org/wiki/How_to_hotswap_UltraBay_devices#Script_for_Ultrabay_eject > to power off the ultrabay. > > Then when I hibernate with the ultrabay powered off, the "Atomic > copy/restore" phase takes 30 seconds and sometimes even a hard lockup > occurs. > Just at the start of "Atomic copy/restore", the drive gets polled (the > light flashes for a second and makes some loading noise) and then > takes the CPU to 100% (I can hear the fan roar) for 30 seconds. > > I'm sure this is related to the fact that the ultrabay is powered off. > There nothing else that triggers this. > > I tried using KDB to interrupt during this phase but that doesn't > work. Only after the operation is complete KDB jumps in, so there is > no way to interrupt whatever is happening. > > I reported this issue to the TuxOnIce developer, and he asked me to > report it to you. Can you please attach the output of dmidecode? Thanks. -- tejun