Re: [RFC][PATCH] PM: Do not create wakeup sysfs files for devices that cannot wakeup
From: Alan Stern <stern@rowland.harvard.edu>
Date: 2011-01-14 20:08:26
On Fri, 14 Jan 2011, Rafael J. Wysocki wrote:
From: Rafael J. Wysocki <redacted> Currently, wakeup sysfs attributes are created for all devices, regardless of whether or not they are wakeup-capable. This is excessive and complicates wakeup device identification from user space (i.e. to identify wakeup-capable devices user space has to read /sys/devices/.../power/wakeup for all devices and see if these files are not empty). Fix this issue by avoiding to create wakeup sysfs files for devices that cannot wake up the system from sleep states (i.e. whose power.can_wakeup flags are unset during registration) and modify device_set_wakeup_capable() so that it adds (or removes) the relevant sysfs attributes if a device's wakeup capability status is changed.
...
quoted hunk
static struct attribute *runtime_attrs[] = { -#ifndef CONFIG_PM_ADVANCED_DEBUG +#ifdef CONFIG_PM_RUNTIME &dev_attr_runtime_status.attr, -#endif &dev_attr_control.attr, &dev_attr_runtime_suspended_time.attr, &dev_attr_runtime_active_time.attr, &dev_attr_autosuspend_delay_ms.attr, +#ifdef CONFIG_PM_ADVANCED_DEBUG + &dev_attr_runtime_usage.attr, + &dev_attr_runtime_active_kids.attr, + &dev_attr_runtime_enabled.attr, +#endif +#endif /* CONFIG_PM_RUNTIME */ NULL, }; static struct attribute_group pm_runtime_attr_group = {@@ -480,35 +485,49 @@ int dpm_sysfs_add(struct device *dev) int rc; rc = sysfs_create_group(&dev->kobj, &pm_attr_group); - if (rc == 0 && !dev->power.no_callbacks) { + if (rc) + return rc; + + if (pm_runtime_callbacks_present(dev)) { rc = sysfs_merge_group(&dev->kobj, &pm_runtime_attr_group); if (rc) - sysfs_remove_group(&dev->kobj, &pm_attr_group); + goto err_out; + }
One thing I don't like about this change. The original code defines runtime_status, runtime_usage, runtime_active_kids, and runtime_enabled whenever PM_RUNTIME and PM_ADVANCED_DEBUG are set. Now you don't define these attributes if runtime callbacks aren't present, even though they would still make sense. Maybe you should have two runtime-related attribute groups, where one depends on callbacks_present and the other doesn't. Apart from that, it all seems reasonable. But you should change the description of the wakeup attribute in Documentation/power/devices.txt; it says that the attribute file is present but empty if the device is not wakeup-capable. Alan Stern