PM: Do not hold dpm_list_mtx while disabling/enabling nonboot CPUs

We shouldn't hold dpm_list_mtx while executing
[disable|enable]_nonboot_cpus(), because theoretically this may lead
to a deadlock as shown by the following example (provided by Johannes
Berg):

CPU 3 CPU 2 CPU 1
suspend/hibernate
something:
rtnl_lock() device_pm_lock()
-> mutex_lock(&dpm_list_mtx)

mutex_lock(&dpm_list_mtx)

linkwatch_work
-> rtnl_lock()
disable_nonboot_cpus()
-> flush CPU 3 workqueue

Fortunately, device drivers are supposed to stop any activities that
might lead to the registration of new device objects way before
disable_nonboot_cpus() is called, so it shouldn't be necessary to
hold dpm_list_mtx over the entire late part of device suspend and
early part of device resume.

Thus, during the late suspend and the early resume of devices acquire
dpm_list_mtx only when dpm_list is going to be traversed and release
it right after that.

This patch is reported to fix the regressions tracked as
http://bugzilla.kernel.org/show_bug.cgi?id=13245.

Signed-off-by: Rafael J. Wysocki <[e-mail]>
Acked-by: Alan Stern <[e-mail]>
Reported-by: Miles Lane <[e-mail]>
Tested-by: Ming Lei <[e-mail]>

drivers/base/power/main.c

···
 {
 	struct device *dev;
 
+	mutex_lock(&dpm_list_mtx);
 	list_for_each_entry(dev, &dpm_list, power.entry)
 		if (dev->power.status > DPM_OFF) {
 			int error;
···
 			if (error)
 				pm_dev_err(dev, state, " early", error);
 		}
+	mutex_unlock(&dpm_list_mtx);
 }
 
 /**
···
 	int error = 0;
 
 	suspend_device_irqs();
+	mutex_lock(&dpm_list_mtx);
 	list_for_each_entry_reverse(dev, &dpm_list, power.entry) {
 		error = suspend_device_noirq(dev, state);
 		if (error) {
···
 		}
 		dev->power.status = DPM_OFF_IRQ;
 	}
+	mutex_unlock(&dpm_list_mtx);
 	if (error)
 		device_power_up(resume_event(state));
 	return error;

kernel/kexec.c

···
 		error = device_suspend(PMSG_FREEZE);
 		if (error)
 			goto Resume_console;
-		device_pm_lock();
 		/* At this point, device_suspend() has been called,
 		 * but *not* device_power_down(). We *must*
 		 * device_power_down() now.  Otherwise, drivers for
···
 		enable_nonboot_cpus();
 		device_power_up(PMSG_RESTORE);
  Resume_devices:
-		device_pm_unlock();
 		device_resume(PMSG_RESTORE);
  Resume_console:
 		resume_console();

kernel/power/disk.c

···
 	if (error)
 		return error;
 
-	device_pm_lock();
-
 	/* At this point, device_suspend() has been called, but *not*
 	 * device_power_down(). We *must* call device_power_down() now.
 	 * Otherwise, drivers for some devices (e.g. interrupt controllers)
···
 	if (error) {
 		printk(KERN_ERR "PM: Some devices failed to power down, "
 			"aborting hibernation\n");
-		goto Unlock;
+		return error;
 	}
 
 	error = platform_pre_snapshot(platform_mode);
···
 
 	device_power_up(in_suspend ?
 		(error ? PMSG_RECOVER : PMSG_THAW) : PMSG_RESTORE);
-
- Unlock:
-	device_pm_unlock();
 
 	return error;
 }
···
 {
 	int error;
 
-	device_pm_lock();
-
 	error = device_power_down(PMSG_QUIESCE);
 	if (error) {
 		printk(KERN_ERR "PM: Some devices failed to power down, "
 			"aborting resume\n");
-		goto Unlock;
+		return error;
 	}
 
 	error = platform_pre_restore(platform_mode);
···
 	platform_restore_cleanup(platform_mode);
 
 	device_power_up(PMSG_RECOVER);
-
- Unlock:
-	device_pm_unlock();
 
 	return error;
 }
···
 		goto Resume_devices;
 	}
 
-	device_pm_lock();
-
 	error = device_power_down(PMSG_HIBERNATE);
 	if (error)
-		goto Unlock;
+		goto Resume_devices;
 
 	error = hibernation_ops->prepare();
 	if (error)
···
 	hibernation_ops->finish();
 
 	device_power_up(PMSG_RESTORE);
-
- Unlock:
-	device_pm_unlock();
 
  Resume_devices:
 	entering_platform_hibernation = false;

kernel/power/main.c

···
 {
 	int error;
 
-	device_pm_lock();
-
 	if (suspend_ops->prepare) {
 		error = suspend_ops->prepare();
 		if (error)
-			goto Done;
+			return error;
 	}
 
 	error = device_power_down(PMSG_SUSPEND);
···
  Platfrom_finish:
 	if (suspend_ops->finish)
 		suspend_ops->finish();
-
- Done:
-	device_pm_unlock();
 
 	return error;
 }