PM / Domains: defer dev_pm_domain_set() until genpd->attach_dev succeeds if present

commit 975e83cf upstream.

If the genpd->attach_dev or genpd->power_on fails, genpd_dev_pm_attach
may return -EPROBE_DEFER initially. However genpd_alloc_dev_data sets
the PM domain for the device unconditionally.

When subsequent attempts are made to call genpd_dev_pm_attach, it may
return -EEXISTS checking dev->pm_domain without re-attempting to call
attach_dev or power_on.

platform_drv_probe then attempts to call drv->probe as the return value
-EEXIST != -EPROBE_DEFER, which may end up in a situation where the
device is accessed without it's power domain switched on.

Fixes: f104e1e5 (PM / Domains: Re-order initialization of generic_pm_domain_data)
Signed-off-by: Sudeep Holla <sudeep.holla@arm.com>
Acked-by: Ulf Hansson <ulf.hansson@linaro.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

drivers/base/power/domain.c

@@ -1188,7 +1188,6 @@ static struct generic_pm_domain_data *genpd_alloc_dev_data(struct device *dev,
 	}
 
 	dev->power.subsys_data->domain_data = &gpd_data->base;
-	dev->pm_domain = &genpd->domain;
 
 	spin_unlock_irq(&dev->power.lock);
 
@@ -1207,7 +1206,6 @@ static void genpd_free_dev_data(struct device *dev,
 {
 	spin_lock_irq(&dev->power.lock);
 
-	dev->pm_domain = NULL;
 	dev->power.subsys_data->domain_data = NULL;
 
 	spin_unlock_irq(&dev->power.lock);
@@ -1248,6 +1246,8 @@ int __pm_genpd_add_device(struct generic_pm_domain *genpd, struct device *dev,
 	if (ret)
 		goto out;
 
+	dev->pm_domain = &genpd->domain;
+
 	genpd->device_count++;
 	genpd->max_off_time_changed = true;
 
@@ -1299,6 +1299,8 @@ int pm_genpd_remove_device(struct generic_pm_domain *genpd,
 	if (genpd->detach_dev)
 		genpd->detach_dev(genpd, dev);
 
+	dev->pm_domain = NULL;
+
 	list_del_init(&pdd->list_node);
 
 	mutex_unlock(&genpd->lock);