Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1753565Ab0AWXmk (ORCPT ); Sat, 23 Jan 2010 18:42:40 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1753239Ab0AWXmJ (ORCPT ); Sat, 23 Jan 2010 18:42:09 -0500 Received: from ogre.sisk.pl ([217.79.144.158]:43514 "EHLO ogre.sisk.pl" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752888Ab0AWXmB (ORCPT ); Sat, 23 Jan 2010 18:42:01 -0500 From: "Rafael J. Wysocki" To: LKML Subject: [PATCH 5/8] PM: Start asynchronous resume threads upfront Date: Sun, 24 Jan 2010 00:38:56 +0100 User-Agent: KMail/1.12.3 (Linux/2.6.33-rc4-rjw; KDE/4.3.3; x86_64; ; ) Cc: Alan Stern , Linus Torvalds , Linux PCI , pm list , linux-usb@vger.kernel.org, Greg KH , Jesse Barnes , James Bottomley , Linux SCSI , Arjan van de Ven , ACPI Devel Maling List , Len Brown , Nigel Cunningham References: <201001240033.34253.rjw@sisk.pl> In-Reply-To: <201001240033.34253.rjw@sisk.pl> MIME-Version: 1.0 Content-Type: Text/Plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Message-Id: <201001240038.56372.rjw@sisk.pl> Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 4106 Lines: 126 From: Rafael J. Wysocki It has been shown by testing that total device resume time can be reduced significantly (by as much as 50% or more) if the async threads executing some devices' resume routines are all started before the main resume thread starts to handle the "synchronous" devices. This is a consequence of the fact that the slowest devices tend to be located at the end of dpm_list, so their resume routines are started very late. Consequently, they have to wait for all the preceding "synchronous" devices before their resume routines can be started by the main resume thread, even if they are "asynchronous". By starting their async threads upfront we effectively move those devices towards the beginning of dpm_list, without breaking their ordering with respect to their parents and children. As a result, their resume routines are started much earlier and we are able to save much more device resume time this way. Signed-off-by: Rafael J. Wysocki --- drivers/base/power/main.c | 43 ++++++++++++++++++++++++------------------- 1 file changed, 24 insertions(+), 19 deletions(-) Index: linux-2.6/drivers/base/power/main.c =================================================================== --- linux-2.6.orig/drivers/base/power/main.c +++ linux-2.6/drivers/base/power/main.c @@ -495,12 +495,12 @@ static int legacy_resume(struct device * } /** - * __device_resume - Execute "resume" callbacks for given device. + * device_resume - Execute "resume" callbacks for given device. * @dev: Device to handle. * @state: PM transition of the system being carried out. * @async: If true, the device is being resumed asynchronously. */ -static int __device_resume(struct device *dev, pm_message_t state, bool async) +static int device_resume(struct device *dev, pm_message_t state, bool async) { int error = 0; @@ -510,6 +510,8 @@ static int __device_resume(struct device dpm_wait(dev->parent, async); down(&dev->sem); + dev->power.status = DPM_RESUMING; + if (dev->bus) { if (dev->bus->pm) { pm_dev_dbg(dev, state, ""); @@ -553,24 +555,16 @@ static void async_resume(void *data, asy struct device *dev = (struct device *)data; int error; - error = __device_resume(dev, pm_transition, true); + error = device_resume(dev, pm_transition, true); if (error) pm_dev_err(dev, pm_transition, " async", error); put_device(dev); } -static int device_resume(struct device *dev) +static bool is_async(struct device *dev) { - INIT_COMPLETION(dev->power.completion); - - if (pm_async_enabled && dev->power.async_suspend - && !pm_trace_is_enabled()) { - get_device(dev); - async_schedule(async_resume, dev); - return 0; - } - - return __device_resume(dev, pm_transition, false); + return dev->power.async_suspend && pm_async_enabled + && !pm_trace_is_enabled(); } /** @@ -583,22 +577,33 @@ static int device_resume(struct device * static void dpm_resume(pm_message_t state) { struct list_head list; + struct device *dev; ktime_t starttime = ktime_get(); INIT_LIST_HEAD(&list); mutex_lock(&dpm_list_mtx); pm_transition = state; - while (!list_empty(&dpm_list)) { - struct device *dev = to_device(dpm_list.next); + list_for_each_entry(dev, &dpm_list, power.entry) { + if (dev->power.status < DPM_OFF) + continue; + + INIT_COMPLETION(dev->power.completion); + if (is_async(dev)) { + get_device(dev); + async_schedule(async_resume, dev); + } + } + + while (!list_empty(&dpm_list)) { + dev = to_device(dpm_list.next); get_device(dev); - if (dev->power.status >= DPM_OFF) { + if (dev->power.status >= DPM_OFF && !is_async(dev)) { int error; - dev->power.status = DPM_RESUMING; mutex_unlock(&dpm_list_mtx); - error = device_resume(dev); + error = device_resume(dev, state, false); mutex_lock(&dpm_list_mtx); if (error) -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/