From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1757393Ab0ANWdg (ORCPT ); Thu, 14 Jan 2010 17:33:36 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1756742Ab0ANWdb (ORCPT ); Thu, 14 Jan 2010 17:33:31 -0500 Received: from kroah.org ([198.145.64.141]:60642 "EHLO coco.kroah.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1756772Ab0ANW2v (ORCPT ); Thu, 14 Jan 2010 17:28:51 -0500 From: Greg Kroah-Hartman To: linux-kernel@vger.kernel.org, stable@kernel.org, stable-review@kernel.org Cc: torvalds@linux-foundation.org, akpm@linux-foundation.org, Ian Campbell , Greg Kroah-Hartman Subject: [PATCH 38/52] xen: fix hang on suspend. Date: Thu, 14 Jan 2010 14:27:17 -0800 Message-Id: <1263508051-7868-38-git-send-email-gregkh@suse.de> X-Mailer: git-send-email 1.6.6 In-Reply-To: <1263508051-7868-37-git-send-email-gregkh@suse.de> References: <20100114222551.GA7839@suse.de> <1263508051-7868-1-git-send-email-gregkh@suse.de> <1263508051-7868-2-git-send-email-gregkh@suse.de> <1263508051-7868-3-git-send-email-gregkh@suse.de> <1263508051-7868-4-git-send-email-gregkh@suse.de> <1263508051-7868-5-git-send-email-gregkh@suse.de> <1263508051-7868-6-git-send-email-gregkh@suse.de> <1263508051-7868-7-git-send-email-gregkh@suse.de> <1263508051-7868-8-git-send-email-gregkh@suse.de> <1263508051-7868-9-git-send-email-gregkh@suse.de> <1263508051-7868-10-git-send-email-gregkh@suse.de> <1263508051-7868-11-git-send-email-gregkh@suse.de> <1263508051-7868-12-git-send-email-gregkh@suse.de> <1263508051-7868-13-git-send-email-gregkh@suse.de> <1263508051-7868-14-git-send-email-gregkh@suse.de> <1263508051-7868-15-git-send-email-gregkh@suse.de> <1263508051-7868-16-git-send-email-gregkh@suse.de> <1263508051-7868-17-git-send-email-gregkh@suse.de> <1263508051-7868-18-git-send-email-gregkh@suse.de> <1263508051-7868-19-git-send-email-gregkh@suse.de> <1263508051-7868-20-git-send-email-gregkh@suse.de> <1263508051-7868-21-git-send-email-gregkh@suse.de> <1263508051-7868-22-git-send-email-gregkh@suse.de> <1263508051-7868-23-git-send-email-gregkh@suse.de> <1263508051-7868-24-git-send-email-gregkh@suse.de> <1263508051-7868-25-git-send-email-gregkh@suse.de> <1263508051-7868-26-git-send-email-gregkh@suse.de> <1263508051-7868-27-git-send-email-gregkh@suse.de> <1263508051-7868-28-git-send-email-gregkh@suse.de> <1263508051-7868-29-git-send-email-gregkh@suse.de> <1263508051-7868-30-git-send-email-gregkh@suse.de> <1263508051-7868-31-git-send-email-gregkh@suse.de> <1263508051-7868-32-git-send-email-gregkh@suse.de> <1263508051-7868-33-git-send-email-gregkh@suse.de> <1263508051-7868-34-git-send-email-gregkh@suse.de> <1263508051-7868-35-git-send-email-gregkh@suse.de> <1263508051-7868-36-git-send-email-gregkh@suse.de> <1263508051-7868-37-git-send-email-gregkh@suse.de> Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org From: Ian Campbell commit c5cae661d6cf808b6984762f763261adf35f3eb7 upstream. In 65f63384 "xen: improve error handling in do_suspend" I said: - xs_suspend()/xs_resume() and dpm_suspend_noirq()/dpm_resume_noirq() were not nested in the obvious way. and changed the ordering of the calls as so: BEFORE AFTER xs_suspend dpm_suspend_noirq dpm_suspend_noirq xs_suspend *SUSPEND* *SUSPEND* dpm_resume_noirq dpm_resume_noirq xs_resume xs_resume Clearly this is not an improvement and I was talking rubbish. In particular the new ordering is susceptible to a hang if a xenstore write is in progress at the point at which the suspend kicks in. When the suspend process calls xs_suspend it tries to take the request_mutex but if a write is in progress it could be looping in xenbus_xs.c:read_reply() waiting for something to arrive on &xs_state.reply_list while holding the request_mutex (taken in the caller of read_reply). However if we have done dpm_suspend_noirq before xs_suspend then we won't get any more xenstore interrupts and process_msg() will never be woken up to add anything to the reply_list. Fix this by calling xs_suspend before dpm_suspend_noirq. If dpm_suspend_noirq fails then make sure we go through the xs_suspend_cancel() code path. Signed-off-by: Ian Campbell Acked-by: Jeremy Fitzhardinge Signed-off-by: Greg Kroah-Hartman --- drivers/xen/manage.c | 8 ++++---- 1 files changed, 4 insertions(+), 4 deletions(-) diff --git a/drivers/xen/manage.c b/drivers/xen/manage.c index c499793..5d42d55 100644 --- a/drivers/xen/manage.c +++ b/drivers/xen/manage.c @@ -102,15 +102,15 @@ static void do_suspend(void) goto out_thaw; } + printk(KERN_DEBUG "suspending xenstore...\n"); + xs_suspend(); + err = dpm_suspend_noirq(PMSG_SUSPEND); if (err) { printk(KERN_ERR "dpm_suspend_noirq failed: %d\n", err); goto out_resume; } - printk(KERN_DEBUG "suspending xenstore...\n"); - xs_suspend(); - err = stop_machine(xen_suspend, &cancelled, cpumask_of(0)); dpm_resume_noirq(PMSG_RESUME); @@ -120,13 +120,13 @@ static void do_suspend(void) cancelled = 1; } +out_resume: if (!cancelled) { xen_arch_resume(); xs_resume(); } else xs_suspend_cancel(); -out_resume: dpm_resume_end(PMSG_RESUME); /* Make sure timer events get retriggered on all CPUs */ -- 1.6.6