From: "Jan Beulich" Subject: [PATCH] linux pciback/pcifront: work queue management fixes Date: Thu, 19 Mar 2009 08:16:50 +0000 Message-ID: <49C20D82.76E4.0078.0@novell.com> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: quoted-printable Return-path: Content-Disposition: inline List-Unsubscribe: , List-Post: List-Help: List-Subscribe: , Sender: xen-devel-bounces@lists.xensource.com Errors-To: xen-devel-bounces@lists.xensource.com To: xen-devel@lists.xensource.com List-Id: xen-devel@lists.xenproject.org flush_scheduled_work() only flushes work queued to the global keventd_wq, but pciback is using its own local work queue, so that is what needs to be flushed. Calling cancel_delayed_work() on something never inserted through queue_delayed_work() or schedule_delayed_work() is pointless. Signed-off-by: Jan Beulich --- head-2009-03-18.orig/drivers/xen/pciback/xenbus.c 2009-03-18 = 10:39:32.000000000 +0100 +++ head-2009-03-18/drivers/xen/pciback/xenbus.c 2009-03-19 = 08:53:57.000000000 +0100 @@ -53,10 +53,9 @@ static void pciback_disconnect(struct pc pdev->evtchn_irq =3D INVALID_EVTCHN_IRQ; } =20 - /* If the driver domain started an op, make sure we complete it or - * delete it before releasing the shared memory */ - cancel_delayed_work(&pdev->op_work); - flush_scheduled_work(); + /* If the driver domain started an op, make sure we complete it + * before releasing the shared memory */ + flush_workqueue(pciback_wq); =20 if (pdev->sh_info !=3D NULL) { xenbus_unmap_ring_vfree(pdev->xdev, pdev->sh_area); --- head-2009-03-18.orig/drivers/xen/pcifront/xenbus.c 2009-03-18 = 10:39:32.000000000 +0100 +++ head-2009-03-18/drivers/xen/pcifront/xenbus.c 2009-03-19 = 08:57:29.000000000 +0100 @@ -64,7 +64,6 @@ static void free_pdev(struct pcifront_de pcifront_free_roots(pdev); =20 /*For PCIE_AER error handling job*/ - cancel_delayed_work(&pdev->op_work); flush_scheduled_work(); unbind_from_irqhandler(pdev->evtchn, pdev); =20